The Chile 2015 Illapel Earthquake and

Tsunami 1st Edition Carla Braitenberg

Visit to download the full and correct content document:
https://textbookfull.com/product/the-chile-2015-illapel-earthquake-and-tsunami-1st-edi
tion-carla-braitenberg/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Geodynamics and Earth Tides Observations from Global to

Micro Scale Carla Braitenberg

https://textbookfull.com/product/geodynamics-and-earth-tides-
observations-from-global-to-micro-scale-carla-braitenberg/

High-Performance Computing for Structural Mechanics and

Earthquake/Tsunami Engineering 1st Edition Shinobu
Yoshimura

https://textbookfull.com/product/high-performance-computing-for-
structural-mechanics-and-earthquake-tsunami-engineering-1st-
edition-shinobu-yoshimura/

Evolving Narratives of Hazard and Risk: The Gorkha

Earthquake, Nepal, 2015 1st Edition Louise Bracken

https://textbookfull.com/product/evolving-narratives-of-hazard-
and-risk-the-gorkha-earthquake-nepal-2015-1st-edition-louise-
bracken/

Tsunami and Fukushima Disaster Design for

Reconstruction 1st Edition Rob Roggema

https://textbookfull.com/product/tsunami-and-fukushima-disaster-
design-for-reconstruction-1st-edition-rob-roggema/
Made Me Blush 1st Edition Carla James

https://textbookfull.com/product/made-me-blush-1st-edition-carla-
james/

The Chile Pepper in China 1st Edition Brian R. Dott

https://textbookfull.com/product/the-chile-pepper-in-china-1st-
edition-brian-r-dott/

Global Tsunami Science Past and Future Volume I 1st

Edition Eric L Geist

https://textbookfull.com/product/global-tsunami-science-past-and-
future-volume-i-1st-edition-eric-l-geist/

Australian Mothering: Historical and Sociological 1st

Edition Carla Pascoe Leahy

https://textbookfull.com/product/australian-mothering-historical-
and-sociological-1st-edition-carla-pascoe-leahy/

Fundamentals of Calculus 1st Edition Carla C. Morris

https://textbookfull.com/product/fundamentals-of-calculus-1st-
edition-carla-c-morris/
 Pageoph Topical Volumes

Carla Braitenberg
Alexander B. Rabinovich
Editors

The Chile-2015
(Illapel)
Earthquake
and Tsunami
The Chile-2015
(Illapel) Earthquake
and Tsunami
Edited by
Carla Braitenberg
Alexander B. Rabinovich

Previously published in Pure and Applied Geophysics (PAGEOPH),

Volume 173, No. 2, 4–7, 2016 and Volume 174, No. 1 and 3, 2017
Editors
Carla Braitenberg Alexander B. Rabinovich
Department of Mathematics and Geoscience Institute of Ocean Sciences
University of Trieste Department of Fisheries and Oceans
Trieste Sidney
Italy Canada

and

P.P. Shirshov Institute of Oceanology

Russian Academy of Sciences
Moscow
Russia

ISBN 978-3-319-57821-7 ISBN 978-3-319-57822-4 (eBook)

DOI 10.1007/978-3-319-57822-4

Library of Congress Control Number: 2017940365

Mathematics Subject Classification (2010): 86-XX, 86-06, 86A17, 76-XX

© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned,
specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in
any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by
similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even
in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore
free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true
and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied,
with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Cover illustration taken from Tang, L., Titov, V.V., Moore, C. et al. Pure Appl. Geophys. (2016) 173: 369. doi:10.1007/s00024-015-1226-3

Cover design: deblik, Berlin

Printed on acid-free paper

This book is published under the trade name Birkhäuser (www.birkhauser-science.com)

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

The Peru-Chile subduction zone, where the Nazca Plate subducts beneath the South American Plate,
is one of the most active seismic zones on the Earth. The Great Chilean (Valdivia) Earthquake with the
moment magnitude Mw 9.5 occurred in this zone on 22 May 1960. The earthquake was the strong-
est ever instrumentally recorded; modern estimates demonstrate that almost 25% of the total global
seismic energy released between 1906 and 2005 was related to the Great Chilean earthquake. The 1960
Chilean earthquake was a milestone in modern seismological and seismotectonic research. Many
hundreds of scientific papers were written on this event, initiating tremendous progress in our under-
standing of the earth processes responsible for producing large megathrust earthquakes. Even now,
57 years after the earthquake, studies of this event continue bringing new scientific results. Among
related effects, the earthquake generated strong global free oscillations of the earth, a phenomenon that
had been predicted theoretically at the time. The theory revealed to be correct, and in 1960 the first
experimental proof of the earth oscillations was given from observations from strainmeters, tilt meters
and long period seismographs.
The great international interest to the 1960 event is related to the trans-Pacific catastrophic tsunami
also generated by the earthquake. The great Chilean tsunami of 22 May 1960 struck the nearby coast
of Chile about 15–20 min after the main earthquake shock; 1,655 people were killed by the tsunami,
approximately 3,000 injured and almost 2,000,000 displaced. Propagating across the Pacific Ocean,
approximately 15 hrs later, tsunami waves reached the Hawaiian Islands and killed 61 people. Then
around 22–24 hrs after the earthquake, the tsunami waves arrived at the Pacific coasts of the Philip-
pines, Japan, and Russia; 142 people lost their lives in Japan and 32 more in the Philippines. The high
degree of destruction and loss of life in a number of the Pacific countries located far away from the
source strongly stimulated international cooperation in tsunami research and mitigation, and resulted
in the establishment of the International Coordination Group for the Tsunami Warning System in the
Pacific (ICG/ITSU, at present ICG/PTWS).
The Chilean subduction zone had been relatively silent since 1960, but recently a sequence of great
tsunami generating earthquakes ruptured the megathrust near the 1960 rupture zone. The earthquakes
are the 27 February 2010 (Mw 8.8), 1 April 2014 (Mw 8.2) and 16 September 2015 (Mw 8.3) shocks.
The tsunamis affected the entire Pacific Ocean and reached large amplitudes for certain coastal popu-
lated areas across the Pacific, similar to what occurred during the Chile 1960 event.
During the fifty years that separated the 1960 event with the onset of the recent sequence, the 2004
Sumatra tsunami occurred, which was the most catastrophic tsunami in human history. This event ini-
tiated a major upgrade of the existing network of coastal tide gauges and strongly stimulated devel-
opment of the Deep-ocean Assessment and Reporting of Tsunamis (DART) system and some other
systems of tsunami measurements in the open ocean. As a result, the 2010, 2014 and 2015 tsunamis
have been recorded by hundreds of high-quality coastal and deep-ocean instruments throughout the
entire Pacific Ocean. At the same time, enormous progress in measuring seismicity and strain rate in
space and time along subduction zones was achieved, allowing the determination of deformation rates
and an understanding of rupture processes on the fault system that makes up the entire South Ameri-
can subduction zone.
Naturally, relatively strong seismic events regularly took place in Chile, but none of these events
had a magnitude comparable to the 1960 earthquake. The event of 27 February 2010, a magnitude Mw
8.8 thrust-fault earthquake, occurred near the coast of Central Chile. The epicenter of the earthquake
was located offshore from the Maule region and the earthquake became known as the “Maule earth-
quake”. The source area of the 2010 Chilean earthquake was located immediately to the north of the
rupture zone of the 1960 Great Chilean Earthquake. The 2010 earthquake was one of the most power-
ful earthquakes in human history and the largest in the Southern Hemisphere since 1960. In addition

V
VI Preface

to causing major destruction throughout central Chile and killing several hundred people, the 2010
earthquake generated tremors that were felt in many Argentinean cities, and in southern Peru.
Four years later, on 1 April 2014, a new great thrust earthquake with moment magnitude Mw 8.2
occurred off the coast of northern Chile. The epicenter was located 70 km from Iquique and approx-
imately 1,600 km north from the epicenter of the 2010 Chile earthquake. The earthquake generated
a major tsunami with the maximum run-up of 4.6 m on the coast closest to the earthquake epi-
center. Tsunami waves were observed throughout the entire Pacific Ocean but no noticeable dam-
age was reported except the coast of Chile. Nevertheless, the event attracted considerable scientific
interest.
As the final, third event in a tragic symphony of great Chilean earthquakes, a destructive Mw 8.3
(Mw 8.4 by some estimates) earthquake occurred on 16 September 2015. The epicenter of the earth-
quake was located offshore, 48 km west of Illapel, central Chile, and 480 km north from the epicenter
of the 2010 Maule earthquake. The earthquake generated a major trans-Pacific tsunami that strongly
affected the coast of Chile with the maximum run-up of 13.6 m and was recorded by numerous coastal
tide gauges and open-ocean DART stations. Tsunami waves created severe damage on the near-source
coast of central Chile and killed 15 people. The unprecedented mass evacuation of over one million
people from the coastal zone prevented a much higher death toll.
In response to this tragic event, Pure and Applied Geophysics (PAGEOPH) announced a Topical
Collection of papers “Chile-2015” and opened a “hot-line” for prospective manuscripts. The main focus
was on the 2015 event, but the invitation encompassed previous great Chilean earthquakes and tsuna-
mis. The activity of scientists was very high and a great many papers illustrating the earthquake and
tsunami studies were submitted.
The twenty-two papers are presented here in a volume that describes the Illapel, Chile 2015 earth-
quake (Mw 8.3) and tsunami from various aspects including the source rupture and comparison with
the observations, using seismologic, geodetic, hydrologic and oceanographic data, giving an excellent
documentation of the effects that such a big event generates.
The rupture of the fault is determined with different models and inversion techniques that aim at
explaining near field seismograms and teleseismic worldwide observations as well as geodetic GNSS
and InSar observations. The geodetic observations allowed scientists to determine the complete perma-
nent surface deformation of the 2015 earthquake. Apart from the seismologic observations, amplitudes
of spheroidal mode free oscillations recorded in a network of superconducting gravimeters were also
used to estimate the seismic moment.
The tsunami coastal run-up and tide gauge observations provided additional constraints on the
rupture models based on seismic records. Near real time modeling of tsunami generation was very
effective, and many lives were saved by incorporating real time seismographic records and data from
open-ocean instruments into tsunami forecast models, resulting in only 7 minutes of elapsed time
between initiation of the earthquake and the tsunami warning.
The earthquake was also observed by water level variations of aquifers as far as the Russian plat-
form, and in a transient anomaly of the magnetic field across Brazil. The GNSS networks addressed
the problem of defining the mechanism limiting the extent of rupture during fault break by investigat-
ing interseismic coupling. Highly coupled segments correlate with source regions of the Chilean large
earthquakes. Furthermore, satellite gravity observations provided evidence that topographic highs on
the subducting oceanic plate was a controlling factor in limiting rupture, such that rupture along fault
is greatest between areas subducting oceanic relief.
The collection of papers from researchers and worldwide institutions demonstrates the great atten-
tion that the South American Chilean subduction margin has garnered. The precise documentation of
the broad spectrum of phenomena that are caused or affected by the earthquake rupture is essential to
improve risk reduction of populations exposed to earthquake hazards in this region. Great progress has
been achieved to be able to issue earthquake communication bulletins with accurate information on
the epicenter location and expected damage in real time, and tsunami warnings even ahead of the time
the tsunami wave hits populated areas. The next step is to develop methods that allow issuing warnings
ahead of the arrival of the seismic wave causing destruction.
Preface VII

The editors of this volume would like to strongly acknowledge Clemens Heine (Executive Editor
Mathematics/Applied Sciences, Birkhäuser/Springer-Verlag), Renata Dmowska (PAGEOPH Editor-
in-Chief, Topical issues), Eric Geist (PAGEOPH Editor, Atmospheric and Ocean Sciences) and many
other PAGEOPH editors who strongly supported the preparation of this volume and helped with edit-
ing of the corresponding papers.

Carla Braitenberg
University of Trieste, Italy
Alexander B. Rabinovich
Institute of Ocean Sciences
Department of Fisheries and Oceans
Sidney, Canada
and
P.P. Shirshov Institute of Oceanology
Russian Academy of Sciences
Moscow, Russia
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V
A Review of Source Models of the 2015 Illapel, Chile Earthquake and Insights from
Tsunami Data
Kenji Satake and Mohammad Heidarzadeh
2017, 174, No.1, Pages 1–9, DOI: 10.1007/s00024-016-1450-5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel,
Chile Mw 8.3 ­Earthquake
Lingling Ye, Thorne Lay, Hiroo Kanamori and Keith D. Koper
2016, 173, No.2, Pages 321–332, DOI: 10.1007/s00024-015-1202-y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Rupture Process During the 2015 Illapel, Chile Earthquake: Zigzag-Along-Dip Rupture ­Episodes
Ryo Okuwaki, Yuji Yagi, Rafael Aránguiz, Juan González and Gabriel González
2016, 173, No.4, Pages 1011–1020, DOI: 10.1007/s00024-016-1271-6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Imaging Rupture Process of the 2015 Mw 8.3 Illapel Earthquake Using the US Seismic Array
Bo Li and Abhijit Ghosh
2016, 173, No.7, Pages 2245–2255, DOI: 10.1007/s00024-016-1323-y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Interseismic Coupling, Megathrust Earthquakes and Seismic Swarms Along the Chilean
Subduction Zone (38°–18°S)
M. Métois, C. Vigny and A. Socquet
2016, 173, No.5, Pages 1431–1449, DOI: 10.1007/s00024-016-1280-5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Low-Frequency Centroid Moment Tensor Inversion of the 2015 Illapel Earthquake

from Superconducting-Gravimeter Data
Eliška Zábranová and Ctirad Matyska
2016, 173, No.4, Pages 1021–1027, DOI: 10.1007/s00024-016-1252-9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Coseismic Fault Slip of the September 16, 2015 Mw 8.3 Illapel, Chile Earthquake
Estimated from InSAR Data
Yingfeng Zhang, Guohong Zhang, Eric A. Hetland, Xinjian Shan, Shaoyan Wen and Ronghu Zuo
2016, 173, No.4, Pages 1029–1038, DOI: 10.1007/s00024-016-1266-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Analysis of the Illapel Mw = 8.3 Thrust Earthquake Rupture Zone Using GOCE-Derived G ­ radients
Orlando Álvarez, Agustina Pesce, Mario Gimenez, Andres Folguera, Santiago Soler and
Wenjin Chen
2017, 174, No.1, Pages 47–75, DOI: 10.1007/s00024-016-1376-y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Chile2015: Lévy Flight and Long-Range Correlation Analysis of Earthquake

Magnitudes in Chile
Maria P. Beccar-Varela, Hector Gonzalez-Huizar, Maria C. Mariani, Laura F. Serpa and
Osei K. Tweneboah
2016, 173, No.7, Pages 2257–2266, DOI: 10.1007/s00024-016-1334-8. . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

IX
X Contents

Time-Based Network Analysis Before and After the Mw 8.3 Illapel Earthquake 2015 Chile
Denisse Pastén, Felipe Torres, Benjamín Toledo, Víctor Muñoz, José Rogan and
Juan Alejandro Valdivia
2016, 173, No.7, Pages 2267–2275, DOI: 10.1007/s00024-016-1335-7. . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

Chilean Earthquakes: Aquifer Responses at the Russian Platform

Alina Besedina, Evgeny Vinogradov, Ella Gorbunova and Igor Svintsov
2016, 173, No.4, Pages 1039–1050, DOI: 10.1007/s00024-016-1256-5. . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

Ionospheric Plasma Response to Mw 8.3 Chile Illapel Earthquake on September 16, 2015
C. D. Reddy, Mahesh N. Shrivastava, Gopi K. Seemala, Gabriel González and Juan Carlos Baez
2016, 173, No.5, Pages 1451–1461, DOI: 10.1007/s00024-016-1282-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Remote Sensing of Atmospheric and Ionospheric Signals Prior to the Mw 8.3 Illapel
Earthquake, Chile 2015
Mohammad Reza Mansouri Daneshvar and Friedemann T. Freund
2017, 174, No.1, Pages 11-45, DOI: 10.1007/s00024-016-1366-0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Chile2015: Induced Magnetic Fields on the Z Component by Tsunami Wave Propagation

V. Klausner, T. Almeida, F. C. De Meneses, E. A. Kherani, V. G. Pillat and M. T. A. H. Muella
2016, 173, No.5, Pages 1463–1478, DOI: 10.1007/s00024-016-1279-y. . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

First Report on Seismogenic Magnetic Disturbances over Brazilian Sector

V. Klausner, T. Almeida, F. C. De Meneses, E. A. Kherani, V. G. Pillat, M. T. A. H. Muella and
P. R. Fagundes
2017, 174, No.3, Pages 737–745, DOI: 10.1007/s00024-016-1455-0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

The 16 September 2015 Chile Tsunami from the Post-Tsunami Survey and Numerical
Modeling Perspectives
Rafael Aránguiz, Gabriel González, Juan González, Patricio A. Catalán, Rodrigo Cienfuegos,
Yuji Yagi, Ryo Okuwaki, Luisa Urra, Karla Contreras, Ian Del Rio and Camilo Rojas
2016, 173, No.2, Pages 333–348, DOI: 10.1007/s00024-015-1225-4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

Field Survey of the 2015 Chile Tsunami with Emphasis on Coastal Wetland and
Conservation Areas
Manuel Contreras-López, Patricio Winckler, Ignacio Sepúlveda, Adolfo Andaur-Álvarez,
Fernanda Cortés-Molina, Camila J. Guerrero, Cyntia E. Mizobe, Felipe Igualt,
Wolfgang Breuer, José F. Beyá, Hernán Vergara and Rodrigo Figueroa-Sterquel
2016, 173, No.2, Pages 349–367, DOI: 10.1007/s00024-015-1235-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

A Study of the 2015 Mw 8.3 Illapel Earthquake and Tsunami: Numerical and Analytical
Approaches
Mauricio Fuentes, Sebastián Riquelme, Gavin Hayes, Miguel Medina, Diego Melgar,
Gabriel Vargas, José González and Angelo Villalobos
2016, 173, No.6, Pages 1847–1858, DOI: 10.1007/s00024-016-1305-0. . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Real-Time Assessment of the 16 September 2015 Chile Tsunami and Implications for
Near-Field Forecast
Liujuan Tang, Vasily V. Titov, Christopher Moore and Yong Wei
2016, 173, No.2, Pages 369–387, DOI: 10.1007/s00024-015-1226-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Contents XI

Comparison Between Tsunami Signals Generated by Different Source Models and the
Observed Data of the Illapel 2015 Earthquake
Ignacia Calisto, Matthew Miller and Iván Constanzo
2016, 173, No.4, Pages 1051–1061, DOI: 10.1007/s00024-016-1253-8. . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

Tsunami Characteristics Along the Peru–Chile Trench: Analysis of the 2015 Mw8.3 Illapel,
the 2014 Mw8.2 Iquique and the 2010 Mw8.8 Maule Tsunamis in the Near-field
R. Omira, M. A. Baptista and F. Lisboa
2016, 173, No.4, Pages 1063–1077, DOI: 10.1007/s00024-016-1277-0. . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

Tsunami Penetration in Tidal Rivers, with Observations of the Chile 2015 Tsunami in
Rivers in Japan
Elena Tolkova
2016, 173, No.2, Pages 389–409, DOI: 10.1007/s00024-015-1229-0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Pure Appl. Geophys. 174 (2017), 1–9
Ó 2016 Springer International Publishing
DOI 10.1007/s00024-016-1450-5 Pure and Applied Geophysics

A Review of Source Models of the 2015 Illapel, Chile Earthquake and Insights from Tsunami
Data
KENJI SATAKE1 and MOHAMMAD HEIDARZADEH1,2

Abstract—The 16 September 2015 Illapel, Chile, earthquake 2014 Iquique earthquake (Mw 8.1, Global CMT),
and associated tsunami have been studied by many researchers
from various aspects. This paper reviews studies on the source
and the 16 September 2015 Illapel earthquake (Mw
model of the earthquake and examines tsunami data. The Illapel 8.3, Global CMT). The 2015 Illapel earthquake
earthquake occurred in the source region of previous earthquakes in occurred at 22:54:32 (UT) on 16 September, at
1943 and 1880. The earthquake source was studied using various
31.573°S, 72.674°W at 22.4 km depth, according to
geophysical data, such as near-field seismograms, teleseismic
waveform and backprojection, GPS and InSAR data, and tsunami the United States Geological Survey. The global
waveforms. Most seismological analyses show a duration of centroid moment tensor (CMT) catalogue provided
*100 s with a peak at *50 s. The spatial distribution has some the scaler seismic moment of 3.2 9 1021 Nm. Its
variety, but they all have the largest slip varying from 5 to 16 m
located at 31°S, 72°W, which is *70 km NW of the epicenter. The focal mechanism solution is strike: 7°, dip: 19°, and
shallow slip seems to be extended to the trench axis. A deeper slip rake: 109°, indicating thrust-type faulting having its
patch was proposed from high-frequency seismic data. A tsunami strike parallel to the trench.
earthquake model with a total duration of 250 s and a third asperity
south of the epicenter is also proposed, but we show that the tsu-
Source regions and recurrence of great earth-
nami data do not support this model. quakes offshore Chile have been studied based on
historical, seismological, and tsunami data (Kelleher
Key words: 2015 Illapel earthquake, tsunami, earthquake
1972; Comte and Pardo 1991; Beck et al. 1998). The
source model, Pacific Ocean, Chilean earthquakes.
2010 Maule earthquake is considered to be a re-
rupture of the 1835 earthquake, which was docu-
mented by Darwin during his voyage on Beagle
(Madariaga et al. 2010). The rupture length of the
1. Introduction
2014 Iquiue earthquake was about 200 km, and it was
only a part of that of the previous great earthquake of
Many great earthquakes repeatedly occur off-
1877 (Schurr et al. 2014; Gusman et al. 2015; Lay
shore Chile, where Nazca plate subducts beneath the
et al. 2014).
South American plate (Fig. 1). This seismic zone
In the source region of the 2015 Illapel
hosted the largest earthquake ever recorded, the 22
earthquake, similar earthquakes occurred in 1943
May 1960 Valdivia earthquake (Mw 9.5, Kanamori
and 1880 and a larger earthquake in 1730 (Fig. 1,
1977) that occurred in southern Chile. In northern
Ruiz et al. 2016). The 2015 Illapel earthquake is
and central Chile, three great tsunamigenic earth-
considered to be a re-rupture of the 1943 earth-
quakes occurred in the last decade (Fig. 1): the 27
quake (Mw 7.9, Beck et al. 1998). Tsunami from
February 2010 Maule earthquake (Mw 8.8, Global
the 1943 earthquake was recorded in Japan with
CMT) with more than 500 deaths as a result of both
height of 10 cm in Hanasaki and 25 cm in
the earthquake and associated tsunami, the 1 April
Kushimoto (Hatori 1968; Watanabe 1998). To the
north of the 2015 source region, the tsunamigenic
Atacama earthquake (Ms 8.3) occurred in 1922
1
Earthquake Research Institute, The University of Tokyo,
(Beck et al. 1998) as well as in 1819. The region
Tokyo, Japan. E-mail: satake@eri.u-tokyo.ac.jp south of the 2015 source was ruptured by the 1971
2
Port and Airport Research Institute, Yokosuka, Japan.

1 Reprinted from the journal

 K. Satake and M. Heidarzadeh Pure Appl. Geophys.

80˚W 75˚W 70˚W 65˚W 2. Tsunami Data

15˚S

The great earthquakes offshore Chile have gen-

1784
1868
2014/4/1

2001
Iquique erated tsunamis, which caused damage not only on
Mw 8.1 the Chilean coast but also across the Pacific Ocean.
20˚S
2014 2007 1995 The 1960 Chilean earthquake and tsunami killed
1768

1877

2000 people on the Chilean coast (Atwater et al.

1999). The tsunami caused 61 and 142 fatalities in
Hawaii and Japan, respectively (Atwater et al. 1999;
South Watanabe 1998). The tsunami heights from the 2010
America
plate Maule earthquake were mostly up to 15 m on the
25˚S
Nazca Chilean coast, and the total fatalities were 156. The
plate tsunami height was \2 m on the Japanese coast, but
caused some property damage (Fujii and Satake
1922
1819

2015/9/16 2013).
Illapel The 2015 Illapel earthquake generated local and
30˚S Mw 8.3 trans-Pacific tsunami (Aránguiz et al. 2016; Contr-
1880
19431971 1985
2015

eras-López et al. 2016; Tang et al. 2016; Zaytsev

et al. 2016). Along the Chilean coast, the Pacific
1822
1730

Tsunami Warning Center (PTWC) and National

1906

2010/2/27 Hydrographic and Oceanic Service (SHOA) issued

Maule tsunami threat messages 7 and 8 min following the
2010

35˚S Mw 8.8
earthquake, respectively (Aránguiz et al. 2016). Such
1751
1835

prompt messages and evacuation helped to minimize

tsunami fatalities (eight deaths according to Aránguiz
et al. 2016).
1837
1737

1960

The post-tsunami surveys measured tsunami

40˚S
Figure 1
Focal mechanism solutions of Global CMT project for shallow
(depth \100 km) earthquakes between 2010 and 2015. Those in
red, purple, and yellow colors are earthquakes that occurred within
1 month of the three great earthquakes (2010 Maule, 2014 Iquique,
and 2015 Illapel). Source zones of historical earthquakes are shown
in bars (Kelleher 1972; Comte and Pardo 1991; Beck et al. 1998;
Schurr et al. 2014)
(M 7.8, USGS) and 1985 (M 8.0, USGS) Val-
paraı́so earthquakes and by the 1906 great
earthquake (Ms 8.4) (Ye et al. 2016).
Interseismic coupling along the Chilean plate
boundary has been inferred from dense GPS net-
works, operated by Chilean collaborations with
American, French, and German groups. High- and
low-coupling zones have been identified. The source
regions of recent large earthquakes match with highly
coupled segments (e.g., Moreno et al. 2010; Schurr
et al. 2014; Métois et al. 2016).
heights on the Chilean coast (Aránguiz et al. 2016;
Contreras-López et al. 2016). The maximum runup
height was reported as 10.8 m at Totoral (30.37°S) by
Aránguiz et al. (2016), whereas it was 13.6 m at La
Cebada (30.97°S) by Contreras-López et al. (2016).
Except for such anomalous locations, the tsunami
heights were up to 9 m on the coast between 29°S
and 32°S, and smaller further south and north
(Fig. 2).
The tsunami was also recorded on coastal tide
gauges (Aránguiz et al. 2016; Heidarzadeh et al.
2016). The earliest tsunami arrival of *15 min with
zero-to-peak amplitude of \2 m was recorded at
Pichidangui tide gauge station, just south of the epi-
center. To the north, at Coquimbo tide gauge station,
the first arrival was at 23 min with *1 m amplitude,
but the largest tsunami amplitude of 4.7 m was
recorded *1.5 h after the earthquake. Omira et al.
(2016) made a comparison of the near-field tsunami
records from the three earthquakes.

Reprinted from the journal 2

Vol. 174, (2017) A Review of Source Models of the 2015 Illapel, Chile Earthquake and Insights

San Felix Trench axis Chanaral

Caldera
Mw 8.4

SOUTH AMERICAN PLATE

2015/9/16
Huasco

PACIFIC Coquimbo
OCEAN
Latitude (deg)

Slip
NAZCA PLATE
Pichidangui
Quintero
Valparaiso
San Antonio
Juan Fernandez

CHILE
Bucalemu
Tide gauge station Constitucion
Tide gauge amplitude
(Heidarzadeh et al., 2016)
Runup survey Quiriquina
(Contreras-Lopez et al., 2016) Coronel
Talcahuano
Runup survey (Aranguiz et al., 2016) Lebu

Tsunami heights (m)

Figure 2
Maximum tide gauge amplitudes of the 2015 Illapel tsunami (pink columns after Heidarzadeh et al. 2016) along with the tsunami heights
reported by Aránguiz et al. (2016) (green circles) and Contreras-López et al. (2016) (black circles). The slip model is from Heidarzadeh et al.
(2016)

While the tide gauge tsunami amplitudes vary in
the range 1–2 m with a maximum value of 4.7 m,
runup heights are in the range 3–6 m with a maxi-
mum value of *11–14 m (Fig. 2). Tsunami runup
heights are roughly up to three times of the tide gauge
amplitudes along the Chilean coast.
According to NOAA Global Historical Tsunami
Database (doi:10.7289/V5PN93H7, https://www.
ngdc.noaa.gov/hazard/tsu_db.shtml, accessed on 24
October 2016), the tsunami heights across the Pacific
Ocean were as follows: 1.37 m in Marquesas Islands,
0.83 m in Hawaii (Hilo), 0.52 m in New Zealand
(Chatham), and 0.10 m in Australia (Port Kembla). In
Japan, Japan Meteorological Agency issued tsunami
advisory at 18:00 on September 17 (UT), approxi-
mately 19 h after the earthquake and 3 h before the
first tsunami arrival to Japan, with expected tsunami
heights of \1 m. The largest observed tsunami
amplitude was 0.78 m on Kuji GPS buoy.
The tsunami was also recorded in deep ocean
on Deep-ocean Assessment and Reporting of Tsu-
namis (DART) buoys. Figure 3 compares the
tsunami waveforms recorded at four DART stations
from the recent three earthquakes. At station
32412, offshore Peru, the 2015 tsunami arrived at
*2.7 h following the earthquake. The 2015 tsu-
nami amplitude was *10 cm, about the same as
tsunami from the 2014 Iquique earthquake and
about a half of that from the 2010 Maule earth-
quake. At stations 43412 off Mexico and 46409 in
the Aleutians, the 2015 tsunami arrived at 9 and
17 h after the earthquake, respectively, and the
amplitude was a few cm, much smaller than the
2010 tsunami and slightly smaller than the 2014

3 Reprinted from the journal

 K. Satake and M. Heidarzadeh Pure Appl. Geophys.

2015 Illapel 2014 Iquique 32412 43412

2010 Maule

46409 52402
Sea level (cm)

Time (hr)

46409

52402 Mw 8.2
43412 2014 Iquique

Mw 8.4
2015 Illapel 32412
Mw 8.8
2010 Maule

Figure 3
Tsunami waveforms from the recent three great earthquakes recorded at four DART stations. The time on the horizontal axis is from the
earthquake origin time only for the 2015 Illapel tsunami (the blue waveforms). The waveforms for the other two tsunamis (red and black
waveforms) are shifted in time to align the first peaks

tsunami. At station 52402 near Saipan in the 3. Slip Distribution

western Pacific Ocean, the 2015 tsunami arrived at
around 20.5 h. The tsunami amplitudes were a few
cm from the 2010 earthquake and almost noise
level for the 2014 and 2015 earthquakes. These
tsunami waveforms, combined with seismological
and geodetic data, are used to study the source
process of these earthquakes (Fujii and Satake
2013; Yoshimoto et al. 2016; Gusman et al. 2015;
Heidarzadeh et al. 2015).
The slip distribution of the 2015 Illapel earth-
quake was studied using various observation data,
such as near-field seismic data, teleseismic wave-
forms, geodetic data, and tsunami waveforms.
Heidarzadeh et al. (2016) used teleseismic and
tsunami waveforms to estimate the slip distribution.
They first conducted teleseismic waveform inversion
by assuming various maximum rupture velocities.

Reprinted from the journal 4

Vol. 174, (2017) A Review of Source Models of the 2015 Illapel, Chile Earthquake and Insights
The source time (moment rate) functions and the
waveform fits of observed and synthetic waveforms
are insensitive to the choice of rupture velocity and
they are similar for different rupture velocities. The
source time function indicates rupture duration of
*120 s with the peak at *50 s. However, the spatial
distribution of coseismic slip, the resultant seafloor
deformation, and computed tsunami waveforms is
different for different rupture velocities. By com-
paring with the observed tsunami waveforms, they
concluded that the slip model with assumed rupture
velocity of 1.75 km/s best explains the observed
tsunamis waveforms. This model has a large-slip area
of 80 km along strike and 100 km along dip, and the
peak slip is located at around 31°S, 72°W, approxi-
mately 70 km from the trench axis and *70 km
northwest of the epicenter (Fig. 4a). The average slip
of the large-slip area is 5.0 m and the total seismic
moment is 4.42 9 1021 Nm (Mw 8.4).
Li et al. (2016) also used teleseismic and tsunami
waveforms to estimate slip distribution. Their slip
model is basically similar to that of Heidarzadeh et al.
(2016), but the large-slip area extended in the north–
south direction with a total length of 170 km, and the
peak slip is almost 10 m and located closer to the
trench (Fig. 4b). The total seismic moment is
2.6 9 1021 Nm (Mw 8.2).
Ruiz et al. (2016) used the high-rate GPS data at
15 stations to estimate the coseismic slip distribution.
The result is similar to the slip models obtained from
tsunami and seismic waveform data, with the peak
slip of 7 m located at around 31°S, 72°W (Fig. 4c),
and the total moment of 4.2 9 1021 Nm (Mw 8.3).
Zhang et al. (2016) used SAR Interferometry
(InSAR) data to estimate a similar peak, with reverse
dip-slip and right-lateral strike-slip components of
8.3 and 1.5 m, respectively. The seismic moment is
estimated at 3.3 9 1021 Nm (Mw 8.3).
Tilmann et al. (2016) made a joint inversion of
geodetic and teleseismic backprojection data. Their
slip model also has the peak at 31°S but with a value
of 4.8 m (Fig. 3d). The peak slip is less localized than
in the other models.
Melgar et al. (2016) used near-field seismic,
geodetic, and tsunami data, as well as teleseismic
backprojection data, and showed two areas of large
slip (asperities). They are both around 31°S; the
5 Reprinted from the journal
shallow one is west of 72°W and the deep one is east
of 72°W. The deep asperity extends to 45 km depth
with 10 m peak slip at 30 km depth. The shallow
asperity extends to the trench with *10 m peak slip
at 15 km depth. About 5 m slip occurred for 200 km
along the trench (Fig. 4e). Li and Ghosh (2016) also
applied teleseismic backprojection but mapped three
patches located around, to the northeast of, and to the
northwest of the epicenter. Okuwaki et al. (2016)
made a hybrid inversion of teleseismic waveform and
backprojection data, and found secondary high-fre-
quency sources in the deeper part of the rupture area.
Lee et al. (2016) inverted teleseismic data with
Green’s functions computed with a 3D spectral-ele-
ment method. They showed that the rupture occurred
in two stages. During the first stage with duration of
*100 s, asperities I and II ruptured (see Fig. 4f for
locations). These two asperities are similar to those of
Melgar et al. (2016), but the slips are larger: the peak
slips of asperities I and II are 10, and 16 m, respec-
tively (Fig. 4f). The total seismic moment of the first
stage is 3.8 9 1021 Nm (Mw 8.3). What is unique
about their model is the second stage, which starts at
100 s after the start of the first stage and lasts until
250 s. They propose that the second stage is a slow
rupture at offshore asperity III, south of the epicenter.
The second stage has a peak slip of 6 m and a seismic
moment of 1.7 9 1021 Nm (equivalent to Mw 8.1).
They consider the first stage a typical interplate
earthquake while the second stage a tsunami earth-
quake. The total seismic moment is 5.5 9 1021 Nm
(Mw 8.4) which is close to that of Heidarzadeh et al.
(2016) and other studies.
To summarize various slip models, most seismo-
logical studies report the total duration of *100 s
with a peak at *50 s for the 2015 Illapel earthquake.
The spatial distribution has some variety, but they all
have the largest slip located at 31°S, 72°W, and
*70 km NW of the epicenter, but the slip amount
varies from 5 to 16 m, nearly a factor of thee. The
difference in the largest slip may be due to different
subfault size; it is not the slip on the fault but the
seismic moment that controls the amplitudes of
seismic waves, and the seafloor displacements (slip
and fault area) that controls tsunami amplitudes. The
shallow slip seems to be extended to the trench axis.
A deeper slip patch was proposed from high-
 K. Satake and M. Heidarzadeh Pure Appl. Geophys.

(a) Heidarzadeh (S and T) (c) Ruiz (G )

29° (b) Li (S and T) 29°

30° 30°

31°S 31°S 31°S

32° 32°

73° 72°W 71°

73° 72°W 71° 33°
0 5 10
Slip(m) Slip (m) 73° 72°W 71°
0 6 Slip (m)
0 4 8
(e) Melgar (S, G and T)
29° (f) Lee (S)
(d) Tilmann (G and S )
30° 30° 30°

31°S 31°S 31°S

32° 32° 32°

73° 72°W 71°

2 4 6 33°
Slip (m) 33°
73° 72°W 71°
74° 73° 72°W 71° 70° Slip (m)
0 6 12

Figure 4
Six slip distribution models estimated by various data sets. S, T, and G stand for seismic, tsunami, and geodetic data, respectively. References
are a Heidarzadeh et al. (2016), b Li et al. (2016), c Ruiz et al. (2016), d Tilmann et al. (2016), e Melgar et al. (2016), and f Lee et al. (2016)

frequency seismic waves. A model of regular earth-
quake followed by a tsunami earthquake with a total
duration of 250 s and a third asperity south of the
epicenter is proposed by Lee et al. (2016).
Tsunami modeling for different seismic sources
and comparison with observations were carried out
by Calisto et al. (2016) and Fuentes et al. (2016).
4. Insights from Tsunami Data
The locations of maximum tsunami runups
reported by Aránguiz et al. (2016) and Contreras-
López et al. (2016) are at the distances of *130 and
*70 km to the north of epicenter, respectively
(Fig. 2). The largest tsunami heights, both runup and
tide gauge heights, are concentrated at around latitude
of 31°S which corresponds to the zone of maximum
slip (Fig. 2). In addition, tsunami heights to the north
of the epicenter are larger than those to the south
(Fig. 2), possibly confirming the northward propa-
gation of the earthquake rupture from the epicenter as
reported by most seismological studies.
The main tsunami source can be estimated from
the travel time of tsunami waves (arrival times minus
the earthquake origin time) by backward ray tracing
from each station (Fig. 5). The travel time curves
ideally surround the tsunami source area. We used

Reprinted from the journal 6

Vol. 174, (2017) A Review of Source Models of the 2015 Illapel, Chile Earthquake and Insights

does not support tsunami source south of the epi-

center, i.e., offshore asperity III of Lee et al. (2016).
Huasco
Lay et al. (2016) computed tsunamis from the slip
models of Lee et al. (2016), including the offshore
San Felix
Mejillones
asperity III ruptured at 95 s, and found that the
additional source cannot reproduce the observed
Caldera tsunami waveforms at DART and coastal tide gauge
Paposo
32412
32402 stations. Heidarzadeh et al. (2016) also made tsunami
Coquimbo
simulation for such a long fault with a second
asperity south of the epicenter, and concluded that the
Bahia-Mansa
Corral Chañaral model predicted tsunami first arrivals at southern
Antofagasta stations (Pichidangui, Quintero, Valpariso and San
Quiriquina
Coronel Antonio, Constitucion) are not consistent with
Lebu observations.
Constitucion
Talcahuano
Pichidangui 5. Conclusions

Quintero Like many other great earthquakes along the

Valparaiso
Juan
Fernandez
San Antonio
Figure 5
Backward ray tracing of the observed tsunamis. The triangles
indicate the locations of tide gauges. The line contours show
tsunami backward travel times from tide gauge or DART stations.
The same color is used for the line contours, the names, and the
triangles of the same stations, although some stations are located
outside of the map. Black dotted line indicates the inferred tsunami
source. Modified from Heidarzadeh et al. (2016)
30 arc s General Bathymetric Chart of the Oceans
(GEBCO)-2014 bathymetric data, which has resolu-
tion of *1 km, and hence, finer topography or
bathymetry features around the tide gauge stations
cannot be expressed. Romano et al. (2016) estimated
the optimal time shifts of tsunami waveforms recor-
ded on coastal tide gauges as several minutes.
Considering the uncertainties, we can roughly esti-
mate the tsunami source at around 31°S and
extending as far west as 73°W. Note that the eastern
edge of the source is not constrained by the observed
data; hence, the above estimate only limits the off-
shore side of the tsunami source. The main tsunami
source seems to extend to the trench axis. Figure 5
Chilean coast, the 2015 Illapel earthquake occurred in
the same region as some previous events. The pre-
vious earthquakes in this region occurred in 1943 and
1880. The source regions of great earthquakes coin-
cide with regions of high coupling as inferred from
GPS measurements. The source models based on
seismic, tsunami, and geodetic data are mostly simi-
lar: source duration of 100 s with a peak at 50 s and
the largest slip at 31°S, 72°W, approximately 70 km
to the NW of the epicenter. The distribution of tsu-
nami heights is consistent with the slip distribution.
The large-slip area seems to be extended to the trench
axis. Another deeper patch is proposed from high-
frequency seismic waves. A model involving a sec-
ondary tsunami earthquake and a third asperity south
of the epicenter is also proposed, but is not supported
by tsunami data.
Acknowledgements
We thank Alexander Rabinovich (Editor-in-Chief)
for inviting us to contribute this review article.
Tsunami data are downloaded from the sea-level
station monitoring facility of Intergovernmental
Oceanographic Commission (IOC) of UNESCO at:
http://www.ioc-sealevelmonitoring.org/. Some fig-
ures are drafted using the GMT software of Wessel

7 Reprinted from the journal

 K. Satake and M. Heidarzadeh
and Smith (1998). DART data are provided by the
National Oceanic and Atmospheric Administration
(NOAA) of the United States (http://www.ndbc.noaa.
gov/dart.shtml). This work was partially supported by
KAKENHI (16H01838).
REFERENCES
Aránguiz, R., González, G., González, J., Catalan, P. A., Cien-
fuegus, R., Yagi, Y., et al. (2016). The 16 September 2015 Chile
tsunami from the post-tsunami survey and numerical modeling
perspectives. Pure and Applied Geophysics, 173, 333–348.
Atwater, B.F., Cisternas, M.V., Bourgeois, J., Dudley, W.C.,
Hendley II J.W., Stauffer, P.H. (1999). Surviving a tsunami—
lessons from Chile, Hawaii, and Japan (p. 20). U.S. Geological
Survey Circular 1187.
Beck, S., Barrientos, S., Kausel, E., & Reyes, M. (1998). Source
characteristics of historic earthquakes along the central Chile
subduction zone. Journal of South American Earth Sciences, 11,
115–129.
Calisto, I., Miller, M., & Constanzo, I. (2016). Comparison
between tsunami signals generated by different source models
and the observed data of the Illapel 2015 earthquake. Pure and
Applied Geophysics, 173, 1051–1061.
Comte, D., & Pardo, M. (1991). Reappraisal of great historical
earthquakes in the northern Chile and southern Peru seismic
gaps. Natural Hazards, 45, 23–44.
Contreras-López, M., Winckler, P., Sepúlveda, I., et al. (2016).
Field Survey of the 2015 Chile tsunami with emphasis on coastal
wetland and conservation areas. Pure and Applied Geophysics,
173, 349–367.
Fuentes, M., Riquelme, S., Hayes, G., Medina, M., Melgar, D.,
Vargas, G., et al. (2016). A study of the 2015 Mw 8.3 Illapel
earthquake and tsunami: Numerical and analytical approaches.
Pure and Applied Geophysics, 173, 1847–1858.
Fujii, Y., & Satake, K. (2013). Slip distribution and seismic
moment of the 2010 and 1960 Chilean earthquakes inferred from
tsunami waveforms and coastal geodetic data. Pure and Applied
Geophysics, 170, 1493–1509.
Gusman, A. R., Murotani, S., Satake, K., Heidarzadeh, M.,
Gunawan, E., Watada, S., et al. (2015). Fault slip distribution of
the 2014 Iquique, Chile, earthquake estimated from ocean-wide
tsunami waveforms and GPS data. Geophysical Research Let-
ters, 42, 1053–1060.
Hatori, T. (1968). Study on distant tsunamis along the coast of
Japan, Part 2, tsunamis of South American origin. Bulletin of the
Earthquake Research Institute University Tokyo, 46, 345–359.
Heidarzadeh, M., Murotani, S., Satake, K., Ishibe, S., & Gusman,
A. R. (2016). Source model of the 16 September 2015 Illapel,
Chile, Mw 8.4 earthquake based on teleseismic and tsunami data.
Geophysical Research Letters, 43, 643–650.
Heidarzadeh, M., Satake, K., Murotani, S., Gusman, A. R., &
Watada, S. (2015). Deep-water characteristics of the trans-Pacific
tsunami from the 1 April 2014 M w 8.2 Iquique, Chile earth-
quake. Pure and Applied Geophysics, 172, 719–730.
Kanamori, H. (1977). The energy release in great earthquakes.
Journal of Geophysical Research, 82, 2981–2987.
Reprinted from the journal
Pure Appl. Geophys.
Kelleher, J. A. (1972). Rupture zones of large South American
earthquakes and some predictions. Journal of Geophysical
Research, 77, 2087–2103.
Lay, T., Yue, H., Brodsky, E. E., & An, C. (2014). The 1 April
2014 Iquique, Chile, Mw 8.1 earthquake rupture sequence.
Geophysical Research Letters, 41, 3818–3825.
Lee, S.-J., Yeh, T.-Y., Lin, T.-C., Lin, Y.-Y., Song, T.-R. A., &
Huang, B.-S. (2016). Two-stage composite megathrust rupture of
the 2015 Mw8.4 Illapel, Chile, earthquake identified by spectral-
element inversion of teleseismic waves. Geophysical Research
Letters, 43, 4979–4985.
Li, B., & Ghosh, A. (2016). Imaging rupture process of the 2015
Mw 8.3 Illapel earthquake using the US Seismic Array. Pure and
Applied Geophysics, 173, 2245–2255.
Li, L., Lay, T., Cheung, K. F., & Ye, L. (2016). Joint modeling of
teleseismic and tsunami wave observations to constrain the 16
September 2015 Illapel, Chile, Mw 8.3 earthquake rupture pro-
cess. Geophysical Research Letters, 43, 4303–4312.
Madariaga, R., Metois, M., Vigny, C., & Campos, J. (2010).
Central Chile Finally Breaks. Science, 328, 181–182.
Melgar, D., Fan, W., Riquelme, S., Geng, J., Liang, C., Fuentes,
M., et al. (2016). Slip segmentation and slow rupture to the
trench during the 2015, Mw8.3 Illapel, Chile earthquake. Geo-
physical Research Letters, 43, 961–966.
Métois, M., Vigny, C., & Socquet, A. (2016). Interseismic cou-
pling, megathrust earthquakes and seismic swarms along the
Chilean subduction zone (38°–18°S). Pure and Applied Geo-
physics, 173, 1431–1449.
Moreno, M., Rosenau, M., & Onken, O. (2010). 2010 Maule
earthquake slip correlates with pre-seismic locking of Andean
subduction zone. Nature, 467, 198–202.
Okuwaki, R., Yagi, Y., Aránguiz, R., González, J., & González, G.
(2016). Rupture process during the 2015 Illapel, Chile earth-
quake: Zigzag-along-dip rupture episodes. Pure and Applied
Geophysics, 173, 1011–1020.
Omira, R., Baptista, M. A., & Lisboa, F. (2016). Tsunami char-
acteristics along the Peru-Chile trench: Analysis of the 2015
Mw8.3 Illapel, the 2014 Mw8.2 Iquique and the 2010 Mw8.8
Maule tsunamis in the near-field. Pure and Applied Geophysics,
173, 1063–1077.
Romano, F., Piatanesi, A., Lorito, S., Tolomei, C., Atzori, S., &
Murphy, S. (2016). Optimal time alignment of tide-gauge tsu-
nami waveforms in nonlinear inversions: Application to the 2015
Illapel (Chile) earthquake. Geophysical Research Letters, 43,
11226–11235. doi:10.1002/2016GL071310.
Ruiz, S., Klein, E., del Campo, F., Rivera, E., Poli, P., Métois, M.,
et al. (2016). The seismic sequence of the16 September 2015 Mw
8.3 Illapel, Chile, earthquake. Seismological Research Letters,
87, 789–799.
Schurr, B., Asch, G., Hainzl, S., Bedford, J., Hoechner, A., Palo,
M., et al. (2014). Gradual unlocking of plate boundary controlled
initiation of the 2014 Iquique earthquake. Nature, 512, 299–302.
Tang, L., Titov, V. V., Moore, C., & Wei, Y. (2016). Real-time
assessment of the 16 September 2015 Chile tsunami and impli-
cations for near-field forecast. Pure and Applied Geophysics,
173, 369–387.
Tilmann, F., Zhang, Y., Moreno, M., Saul, J., Eckelmann, F., Palo,
M., et al. (2016). The 2015 Illapel earthquake, central Chile: A
type case for a characteristic earthquake? Geophysical Research
Letters, 43, 574–583.
8
Vol. 174, (2017) A Review of Source Models of the 2015 Illapel, Chile Earthquake and Insights

Watanabe, H. (1998). Materials for comprehensive list of
destructive tsunamis in Japan (‘‘Nihon Higai Tsunami Souran’’)
(p. 206). Tokyo: University of Tokyo Press.
Wessel, P., & Smith, W. H. F. (1998). New, improved version of
Generic Mapping Tools released. EOS Transactions, AGU, 79,
579.
Ye, L., Lay, T., Kanamori, H., & Koper, K. D. (2016). Rapidly
estimated seismic source parameters for the 16 September 2015
Illapel, Chile Mw 8.3 earthquake. Pure and Applied Geophysics,
173, 321–332.
Yoshimoto, M., Watada, S., Fujii, Y., & Satake, K. (2016). Source
estimate and tsunami forecast from far-field deep-ocean tsunami
waveforms—The 27 February 2010 Mw 8.8 Maule earthquake.
Geophysical Research Letters, 43, 659–665.
Zaytsev, O., Rabinovich, A. B., & Thomson, R. E. (2016). A
comparative analysis of coastal and open-ocean records of the
great Chilean tsunamis of 2010, 2014 and 2015 off the coast of
Mexico. Pure and Applied Geophysics. doi:10.1007/s00024-016-
1407-8.
Zhang, Y., Zhang, G., Hetland, E. A., Shan, X., Wen, S., & Zuo, R.
(2016). Coseismic fault slip of the September 16, 2015 Mw 8.3
Illapel, Chile earthquake estimated from InSAR data. Pure and
Applied Geophysics, 173, 1029–1038.

(Received November 3, 2016, revised November 30, 2016, accepted December 1, 2016, Published online December 8, 2016)

9 Reprinted from the journal

Pure Appl. Geophys. 173 (2016), 321–332
Ó 2016 Springer International Publishing
DOI 10.1007/s00024-015-1202-y
Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel,
Chile Mw 8.3 Earthquake
LINGLING YE,1 THORNE LAY,2 HIROO KANAMORI,1 and KEITH D. KOPER3
Abstract—On 16 September 2015, a great (Mw 8.3) interplate
thrust earthquake ruptured offshore Illapel, Chile, producing a 4.7-m
local tsunami. The last major rupture in the region was a 1943 MS 7.9
event. Seismic methods for rapidly characterizing the source process,
of value for tsunami warning, were applied. The source moment
tensor could be obtained robustly by W-phase inversion both within
minutes (Chilean researchers had a good solution using regional data
within 5 min) and within an hour using broadband seismic data.
Short-period teleseismic P wave back-projections indicate northward
rupture expansion from the hypocenter at a modest rupture expansion
velocity of 1.5–2.0 km/s. Finite-fault inversions of teleseismic P and
SH waves using that range of rupture velocities and a range of dips
from 16°, consistent with the local slab geometry and some moment
tensor solutions, to 22°, consistent with long-period moment tensor
inversions, indicate a 180- to 240-km bilateral along-strike rupture
zone with larger slip northwest to north of the epicenter (with peak
slip of 7–10 m). Using a shallower fault model dip shifts slip seaward
toward the trench, while a steeper dip moves it closer to the coastline.
Slip separates into two patches as assumed rupture velocity increases.
In all cases, localized *5 m slip extends down-dip below the coast
north of the epicenter. The seismic moment estimates for the range of
faulting parameters considered vary from 3.7 9 1021 Nm (dip 16°) to
2.7 9 1021 Nm (dip 22°), the static stress drop estimates range from
2.6 to 3.5 MPa, and the radiated seismic energy, up to 1 Hz, is about
2.2–3.15 9 1016 J.
Key words: 2015 Illapel earthquake, Chilean seismic gaps,
rupture process, seismic rupture parameters.
This paper is part of the article collection on ‘‘Illapel, Chile,
Earthquake on September 16th, 2015’’.
Electronic supplementary material The online version of this
article (doi:10.1007/s00024-015-1202-y) contains supplementary
material, which is available to authorized users.
1
Seismological Laboratory, California Institute of Technol-
ogy, MS 252-21, Pasadena, CA 91101, USA.
2
Department of Earth and Planetary Sciences, University of
California Santa Cruz, Santa Cruz, CA 95064, USA. E-mail:
tlay@ucsc.edu
3
Department of Geology and Geophysics, University of
Utah, Salt Lake City, UT 84112, USA.
11
Pure and Applied Geophysics
1. Introduction
The subduction zone along Chile was struck by
the third great earthquake in 6 years when the plate
boundary offshore of Illapel ruptured from 30.25°S to
32.5°S in an MW 8.3 event on 16 September 2015
(31.570°S, 71.670°W, 22:54:33 UTC (USGS
National Earthquake Information Center—NEIC:
http://earthquake.usgs.gov/earthquakes/) (Fig. 1).
This region had been identified as a seismic gap
based on the occurrence of prior large earthquakes in
1943, 1880, and 1730 (KELLEHER 1972; NISHENKO
1985; BECK et al. 1998). The region just to the south
(32°S to 34.5°S) ruptured most recently in the 9 July
1971 [MS 7.9, MW 7.8, (ISC-GEM)] and 3 March
1985 [MS 7.8, MW 7.9 (CMT)] Valparaı́so earth-
quakes as well as previously in the great 17 August
1906 (MS 8.4) earthquake (e.g., COMTE et al. 1986;
CHRISTENSEN and RUFF 1986). The region to the north
(26°S to 30.25°S) ruptured in the great 11 November
1922 MS 8.3 earthquake (e.g., BECK et al. 1998)
(actual rupture is likely to be offshore, as large tsu-
nami was produced, so the ISC-GEM location in
Fig. 1a is too far inland). The 1922 rupture zone has
been seismically relatively quiet since then.
The rapid global Centroid-Moment Tensor
(gCMT) solution for this event (http://www.
globalcmt.org/CMTsearch.html) indicates an almost
pure double-couple faulting geometry with strike, /
= 5°, dip, d = 22°, and rake, k = 106°, at a centroid
depth hc = 17.8 km with a centroid location north-
west of the hypocenter (31.22°S, 72.27°W) (Fig. 1).
The centroid time shift tc is 47.9 s and seismic
moment M0 is 2.86 9 1021 Nm (Mw 8.2). The
W-phase solution from CNRS has M0 = 2.68 9 1021
Nm, with a centroid location at 31.02°S and 72.04°W
Reprinted from the journal
 L. Ye et al. Pure Appl. Geophys.

(a) (b)
-25˚
1966(M7.7)

-26˚ 1965(M7.0)
7.4 cm/yr
1918(M7.8) -30˚
1983(M7.4)
-27˚
1939(M7.4)

-28˚
1922(M7.0) 1922(M8.3)

-29˚
-31˚

7.4 cm/yr 1943(8.1)

-30˚
1955(M7.0)

1977(M7.4)
2015 Mw 8.3 1997(M7.1)
-31˚
1943(M8.1) -32˚
1944(M7.0)
-32˚ 1906(M8.2)
1906(M8.2)
1965(M7.4)
1985(M8.0) 1971(M7.8)
-33˚ 1985(M7.5)
1981(M7.2) 1985(M7.4) 1971(M7.8)
1985(M8.0)
1931(M7.2) -33˚
-34˚
7.4 cm/yr 1985(M7.2)

2010(M7.0)
-35˚
-74˚ -73˚ -72˚ -71˚ -70˚ -69˚ -68˚ -67˚ -73˚ -72˚ -71˚
0 10 20 30 40 50 60 70 0 2 4 6 8
Source Depth (km) Coseismic Slip (m)

Figure 1
Maps of seismicity in the source region of the 16 September 2015 Illapel, Chile earthquake with the inset (b) showing a slip distribution
inverted from teleseismic body waves. The USGS-NEIC epicenter is indicated by the red star. The global centroid moment tensor (gCMT)
mechanism is shown with a line to the epicenter in both panels and plotted at the gCMT centroid in (b). gCMT solutions for other large events
are shown at their respective centroid locations, color-coded by depth and scaled proportional to their moments. Mechanisms for the great
1943 and 1922 events are from BECK et al. (1998). Preceding earthquakes larger than magnitude 7.0 from 1900 to 2015 from the ISC-GEM
catalog are shown by the depth-colored circles in (a) and the blue circles in (b) (or focal mechanisms, if known) with each radius scaled by
estimated magnitude (year and magnitude are indicated). The seismicity of the 1997 Coquimbo swarm is outlined in yellow. The black
contours in (a) are the fault locking model of MÉTOIS et al. (2012) for south of -28°, with the dashed curve indicating 50 % locking and the
solid curve 90 % locking. The small green circles in (b) are USGS-NEIC epicenters of aftershocks in the first 2 weeks following the 2015
mainshock. The rectangular area in (b) is the slip distribution estimated by teleseismic P and SH inversion for Vr = 2.0 km/s and dip 22° in
this study (Fig. 4d). The toothed white curves indicate the trench position

with hc = 17.5 km, and best-double-couple / location and size estimates confirm the potential for
= 2.7°, d = 22.3°, and k = 94.3° (http://wphase. large tsunami generation.
unistra.fr/events/illapel_2015/index.html). The The Nazca plate has been underthrusting South
USGS-NEIC W-phase solution has slightly larger America near the 2015 rupture at about 74 mm/year
M0 = 3.19 9 1021 Nm (MW 8.3) and hc = 25.5 km, (DEMETS et al. 2010) for the 72 years since the 6
with a best-double-couple geometry of / = 353°, April 1943 rupture, suggesting that up to 5.3 m of slip
d = 19° and k = 83°. Although the precise seismic deficit may have accumulated since then. Accumu-
moment, dip and source depth vary among these early lation of slip deficit on the Chilean megathrust has
point-source estimates due to intrinsic trade-offs and been measured using the progressively eastward
limitations of long-period determinations, the deflection of GPS stations in Chile for almost two

Reprinted from the journal 12

Vol. 173, (2016) Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel
decades (e.g., NORABUENA et al. 1998; KENDRICK et al.
1999). Several early interpretations modeled GPS
observations with a uniformly totally locked shallow
plate boundary and a significantly deforming back-
arc region (e.g., KENDRICK et al. 2001; KHAZARADZE
and KLOTZ 2003; BROOKS 2003). VIGNY et al. (2009)
conclude that only 40–45 % of the total convergence
rate between the Nazca and South American plates is
causing accumulation of elastic deformation in the
upper plate. Recent interpretations indicate that the
plate boundary slip deficit is a large fraction of the
plate convergence near 31°S, but reduces along
31.5–32°S (MORENO et al. 2010). MÉTOIS et al. (2012)
also argue for variable coupling along the plate
boundary (Fig. 1a), but find nearly 100 % coupling
from 31°S to 33°S at depths from 15 to 45 km on the
megathrust, corresponding to the 2015 rupture area in
Fig. 1b. Weak coupling from 30°S to 31°S is indi-
cated for the region to the north of the rupture where
the July 1997 Coquimbo swarm (Fig. 1) occurred
(e.g., GARDI et al. 2006; HOLTKAMP et al. 2011;
BRODSKY and LAY 2014).
Recent studies of large subduction-zone earth-
quakes clearly demonstrated that although the long-
term strain accumulation patterns can be mapped
quantitatively with GPS, the occurrence of individual
earthquakes exhibit significant complexity in both
spatial and temporal distributions, especially along
subduction zones where events with variable rupture
length have occurred. Given this complexity, when a
large earthquake has occurred, very rapid assessment
of the event is critically important for early hazard
warning of aftershocks and tsunamis, immediate
post-earthquake emergency efforts, and assessment of
regional seismic potential immediately following the
event. With the currently available real-time data and
analysis methods, we can assess the first-order tsu-
nami potential in 5–20 min using long-period seismic
waves and high-rate GPS, and characterize the slip
distribution and the spectral characteristics of the
source within a few hours. In view of this importance,
we report here the primary source parameters esti-
mated from rapid seismological methods for the 16
September 2015 event and discuss the rupture in the
context of the historic earthquake activity and the
plate locking determinations. We examine point-
13
source parameters and tectonic setting to constrain
the faulting geometry, back-projection of teleseismic
P waves to constrain the rupture dimensions and
expansion velocity, and then perform finite-fault
inversions for which we estimate source parameters
for the event.
2. Rapid Seismic Analysis of the 2015 Mw 8.3
Earthquake
2.1. Point-Source Solutions
For tsunami warning purposes, it is desirable to
estimate the earthquake mechanism, seismic moment,
and depth within a few minutes. We applied the
W-phase inversion method (KANAMORI and RIVERA
2008; DUPUTEL et al. 2012) to regional three-compo-
nent ground displacements in the passband 1–5 mHz
from 10 to 11 global seismic stations (22 compo-
nents) within 10° of the epicenter. For source depths
of 17.5–25.5 km this analysis yielded shallow-dip-
ping (13.7°–17.4°) thrust mechanisms with moments
of 3.5–3.4 9 1021 Nm, and could be completed
within five to ten minutes if operated in real time,
as it requires only 200–300 s of recorded signal after
the origin time. Centroid time shifts, tc, of 42–41 s
were found, which are only moderately longer than
the value of 36 s predicted based on the average
relationship found by Duputel et al. (2013) of
tc = 2.59 9 10-6 M1/3 0 (for M0 in Nm). This provides
a rapid determination that the source was not a
tsunami earthquake, as such events have significantly
larger tc shifts relative to the average (DUPUTEL et al.
2013), even while the seismic moment, mechanism,
and location of the 2015 event do suggest significant
tsunamigenesis.
An operational application of regional W-phase
inversion was performed by the Chilean National
Seismological Center of the University of Chile, and
provided a solution within 5 min using data from 22
stations within 12° of the epicenter, with MW 8.1.
This information was automatically transmitted to the
Navy Hydrographic Service in charge of the tsunami
alerts, and the Emergency Office of the Interior
deployed an evacuation plan for the coastal popula-
tion. It is estimated that about 1 million people were
Reprinted from the journal
 L. Ye et al.
evacuated from various exposed regions in central
Chile.
Point-source inversions using global data for the
16 September 2015 Illapel mainshock reported by
routine rapid seismic wave processing have moder-
ately varying thrust plane dip, centroid depth, and
seismic moment. For shallow thrust events the
product M0 9 sin(2d) is well determined by long-
period inversions. However, there is strong trade-off
between dip and seismic moment, such that the
estimated moment is about 30 % larger for a 16° dip
than for a 22° dip. We performed W-phase inversions
using varying global data distributions finding cen-
troid depths of 25.5–30.5 km and dips of 16.1° to
18.5° with centroid locations being dependent on the
data set.
To provide an independent estimate of the dip to
stabilize the trade-off between dip and moment, we
examined the slab geometry models of HAYES and
WALD (2009) and HAYES et al. (2012) which indicate a
relatively planar plate boundary interface dipping at
*16.5° near 31°S. Relative to the trench position, the
USGS-NEIC hypocenter is compatible with a dip of
*16° as well. Thus, there is a range of dip estimates of
from about 16° to 22° quickly available, with the
steeper dip requiring some reduction of dip at shallow
depth on the megathrust to connect to the trench. There
is no clear resolution of shallow curvature from the
seismicity distributions. For our finite-fault inversions
we consider models with uniform dips from 16° to 22°,
recognizing that the seismic moment varies with dip
and that dip may actually vary along the megathrust.
2.2. Back-Projection Analysis
We performed a back-projection of teleseismic
short-period P waves recorded in North America.
This is the best station geometry (Fig. 2a) for back-
projection given the location of the source region and
global station distribution. The broadband P wave
arrivals are very coherent and readily aligned across
North America by multi-station correlation (Fig. 2b).
The aligned traces were then filtered in the passband
0.5–2.0 Hz and back-projected using the procedure of
XU et al. (2009) with fourth-root stacking of beam
power across a constant depth target grid in the
source region. Figure 2c shows the time-integrated
Reprinted from the journal
Pure Appl. Geophys.
beam power at each grid position, demonstrating that
the coherent short-period energy largely originated
north of the epicenter, with two concentrations, the
higher power image forming about 60 km NE of the
epicenter and a second region about 50 km WNW of
the epicenter. The primary bursts of coherent short-
period energy are found in the first 50 s of the rupture
process near the coastal region and there is little
evidence for short-period bursts toward the trench
(Fig. 2d). The spatial extent of the rupture indicated
by these back-projections extends less than 100 km
from the epicenter, which together with the 47.9 s
centroid time for the gCMT inversion suggests a low
overall rupture expansion velocity.
Time snapshots from the back-projection are
shown in Fig. 3, capturing the irregular growth of
the two lobes of short-period energy, first forming to
the WNW from the epicenter in the first 20 s, then to
the NE around 30 s, then with irregular northward
expansion over the next 30 s. Significant coherent
energy is imaged to at least 95 s lag time in the
azimuth to North America. Estimates of rupture
expansion velocity of 1.5–2.0 km/s can be made from
the space–time pattern of these back-projection
images. Similar overall behavior is found in the on-
line back-projections posted by the Incorporated
Research Institutions for Seismology (IRIS) (http://
ds.iris.edu/spud/backprojection/10089385).
Surface wave source time functions display slight
narrowing at azimuths northward from the epicenter,
compatible with a low rupture velocity (1–2 km/s)
along a 5° strike (http://ds.iris.edu/spud/
sourcetimefunction/10090670) or about 30 s range
in duration at an azimuth of 345° with a mean
duration of about 105 s. Assuming unilateral rupture
toward the NNW the latter measurements indicate a
rupture dimension of only about 60 km and a low
rupture velocity of *1 km/s. Some component of
bilateral rupture cannot be precluded, but the overall
long-period source directivity is modest, similar to
the short-period back-projection finiteness.
2.3. Teleseismic Finite-Fault Inversion
Guided by these constraints on the faulting
geometry and rupture dimensions, several finite-fault
slip inversions were performed using a least-squares
14
Vol. 173, (2016) Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel

(a) (b)
90
50˚

75
40˚

Distance (o)
30˚
60

20˚

45
10˚

0˚
30
-120˚ -110˚ -100˚ -90˚ -80˚ -70˚ -60˚ 0 20 40 60 80 100
Time (s)
0.5 0.6 0.7 0.8 0.9 1.0
Correlation Coefficient
(c) (d)
-30˚ -30˚

-31˚ -31˚

-32˚ -32˚

100 km 100 km

-33˚ -33˚
-73˚ -72˚ -71˚ -70˚ -73˚ -72˚ -71˚ -70˚

Beam Amplitude 0 20 40 60 80
Lapse Time (s)

Figure 2
a Location of North American seismic stations used in the back-projection of 0.5–2.0 Hz P wave signals for the 16 September 2015 Illapel
earthquake, color-coded by average correlation coefficient with all other traces. b The aligned, unfiltered P wave data from North America
used in the back-projection. c Map of the time-integrated beam power distribution obtained by back-projection. The star indicates the USGS-
NEIC epicenter, which is the reference position for aligning the first-motions of the P waves. The small circles are USGS-NEIC aftershock
epicenters in the first 4 days following the 2015 mainshock, scaled proportional to magnitude. d Positions of back-projection maxima at
various lapse times from the back-projection image for the North American P wave data. The relative power of peaks scales the circle radii
and the points are color-coded by time into the rupture. An animation of the full back-projection is provided in Supplemental Movie 1

procedure applied to 135 s long recordings of 60
teleseismic broadband P waves and 42 SH waves
(HARTZELL and HEATON 1983; adapted by KIKUCHI and
KANAMORI 1991). In this kinematic inversion we
prescribe a fault strike (5°) and dip (16° or 22°) based
on slab geometry or long-period moment tensors. The
fault is subdivided into 9–10 segments along-dip and
18–19 along-strike with 15 9 15 km subfault dimen-
sions. The subfault source time functions are
parameterized by 6 2.5 s rise time triangles offset

15 Reprinted from the journal

 L. Ye et al. Pure Appl. Geophys.

-20 0 20 40 60 80 100 120 -20 0 20 40 60 80 100 120 -20 0 20 40 60 80 100 120

Time (s) Time (s) Time (s)
-30˚ -30˚ -30˚

-31˚ -31˚ -31˚

-32˚ -32˚ -32˚

-33˚ -33˚ -33˚

-73˚ -72˚ -71˚ -70˚ -73˚ -72˚ -71˚ -70˚ -73˚ -72˚ -71˚ -70˚

-20 0 20 40 60 80 100 120 -20 0 20 40 60 80 100 120 -20 0 20 40 60 80 100 120

Time (s) Time (s) Time (s)
-30˚ -30˚ -30˚

-31˚ -31˚ -31˚

-32˚ -32˚ -32˚

-33˚ -33˚ -33˚

-73˚ -72˚ -71˚ -70˚ -73˚ -72˚ -71˚ -70˚ -73˚ -72˚ -71˚ -70˚

Figure 3
Snapshots from the North American P wave back-projection for the 2015 Illapel earthquake. Each image shows the spatial distribution of
beam power at the lapse time indicated by the red line on the power time history shown above each map. These power peaks can be used to
infer the pattern of rupture expansion with time and the associated effective rupture velocity for the portion of the 0.5- to 2.0-Hz energy that
constructively interferes in the beam image. The fourth-root beam power is shown, so the image is not linearly proportional to wave
amplitudes. Significant beam power is imaged for about 90 s. The small circles are USGS-NEIC aftershock epicenters in the first 4 days
following the 2015 mainshock (red star), scaled proportional to magnitude. The continuous animation is provided in Supplemental Movie 1

by 2.5 s, giving possible total subfault durations of
17.5 s. Rake is allowed to vary for each subfault
subevent. Rupture expansion velocities of 1.5 and
2.0 km/s are considered here, although we considered
a wider range of rupture velocities finding that an
optimal choice is not constrained by waveform misfit,
so we imposed bounds from the back-projections and
R1 time function directivity to settle on this range of
possibilities. The hypocentral depth was 25 km, and
Green’s functions were computed for the local crustal
structure from Crust 1.0 (LASKE et al. 2013) with
addition of a 3 km deep water layer.

Reprinted from the journal 16

Vol. 173, (2016) Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel

Four slip distributions inverted for the 16 Septem- centroid times of the moment rate functions for the
ber 2015 event are shown in Fig. 4. The solution with models vary from 50.5 to 57.3 s.
Vr = 2.0 km/s and dip 22° (Fig. 4d) is placed on the The slip models have modest rake variability, but
regional map in Fig. 1b, as that model is most all share an average rake of about 96°. About
consistent with early reports of large geodetic 80–82 % of the teleseismic P and SH waveform
motions north of the hypocenter. For the shallower power is accounted for by each model, and it is not
dip of 16°, the fault model extends further offshore, sensible to favor a particular fault parameterization
and slip occurs along the up-dip portion of the based on waveform fit. Figure 5 shows an example of
megathrust with slip centroid depths of about 16 km. waveform fits for the model in Fig. 4d. The P waves
For the lower rupture velocity, slip is relatively are particularly well fit for all 4 models.
uniform and as the rupture velocity increases it splits The along-strike distribution of large slip depends
into two patches, with more slip in the north. For the on the assumed rupture velocity and is about 180 km
22° dip cases the slip locates deeper on the fault for Vr = 1.5 km/s and 240 km for Vr = 2.0 km/s. All
closer to the coastline with slip centroid depths of of the models have a concentration of slip to the N or
20–24 km, with two slip patches, the southern one NW that accounts for the modest directivity seen in
locating further up-dip than the northern one. The the R1 source time functions. All of the models also

(a) Vr = 1.5 km/s, δ = 16°, E = 3.1 MPa (b) Vr = 2.0 km/s, δ = 16°, E = 4.0 MPa

-60 11.5 -60 11.5

Distance along dip (km)

Distance along dip (km)

Depth (km)

Depth (km)
-30 19.7 -30 19.7

0 28.0 0 28.0

30 36.3 30 36.3

-120 -90 -60 -30 0 30 60 90 120 150 -120 -90 -60 -30 0 30 60 90 120 150
Distance along Strike (km) Distance along Strike (km)
0 2 4 6 8 10 12
Coseismic Slip(m)
(c) Vr = 1.5 km/s, δ = 22°, E = 2.6 MPa (d) Vr = 2.0 km/s, δ = 22°, E = 2.6 MPa
-60 5.5 -60 5.5
Distance along dip (km)
Distance along dip (km)

-30 16.8 -30 16.8

Depth (km)

Depth (km)

0 28.0 0 28.0

30 39.2 30 39.2

-120 -90 -60 -30 0 30 60 90 120 150 -120 -90 -60 -30 0 30 60 90 120 150
Distance along Strike (km) Distance along Strike (km)

Figure 4
Slip models for the 16 September 2015 Illapel earthquake obtained by least-squares inversion of teleseismic P and SH waves in the passband
0.005–0.9 Hz. a Model for Vr = 1.5 km/s and a dip of 16°. b Model for Vr = 2.0 km/s and a dip of 16°. c Model for Vr = 1.5 km/s and a dip
of 22°. d Model for Vr = 2.0 km/s and a dip of 22°. Waveform fits are very similar for each model; those for d are shown in Fig. 5. Each
model grid has 15 km spacing between subfaults. The black arrows indicate the direction and magnitude of average slip in the fault model
coordinates for each subfault with the color-scale indicating the absolute slip magnitudes. The red star is the USGS-NEIC hypocentral
location. The pink outline indicates the subfaults with estimated moment larger than 15 % of the peak subfault moment used for the simplified
stress drop calculation, Dr0.15

17 Reprinted from the journal

 L. Ye et al. Pure Appl. Geophys.

HRV P 311 PMSA P 311 TUC P 289 HRV SH 1030 HOO SH 1477
(0°, 73 °) (174°, 33 °) (326°, 73 °) (0°, 73 °) (191°, 65 °)

SCHQ P 216 SMAI P 287 BBB P 92 SDV SH 1706 MCQ SH 1176

(2°, 86 °) (177°, 36 °) (329°, 96 °) (1°, 40 °) (206°, 83 °)

BBSR P 337 QSPA P 293 ANMO P 300 SCHQ SH 1035 SNZO SH 765
(6°, 63 °) (180°, 58 °) (330°, 73 °) (2°, 86 °) (223°, 85 °)

SJG P 394 CASY P 229 HKT P 359 HATO SH 1386 RAO SH 587
(6°, 49 °) (180°, 82 °) (337°, 65 °) (3°, 43 °) (237°, 87 °)

DRLN P 333 SBA P 272 RSSD P 263 SJG SH 1173 NIUE SH 346
(9°, 81 °) (191°, 65 °) (337°, 80 °) (6°, 49 °) (249°, 86 °)

NRS P 126 MCQ P 216 JTS P 346 DRLN SH 1033 AFI SH 265
(12°, 94 °) (206°, 83 °) (340°, 43 °) (9°, 81 °) (252°, 91 °)

GRGR P 502 TAU P 84 MGMO P 326 ANWB SH 993 PTCN SH 757

(14°, 44 °) (207°, 97 °) (342°, 71 °) (12°, 49 °) (261°, 51 °)

VAL P 109 MOO P 71 FFC P 155 GRGR SH 1434 XMAS SH 415

(33°, 98 °) (207°, 98 °) (343°, 89 °) (14°, 44 °) (273°, 87 °)

CMLA P 380 SNZO P 266 CCM P 285 CMLA SH 1043 JEDS SH 901
(35°, 81 °) (223°, 85 °) (343°, 71 °) (35°, 81 °) (325°, 88 °)

PAB P 156 RAO P 187 ULM P 235 MESJ SH 642 COR SH 841
(45°, 94 °) (237°, 87 °) (344°, 84 °) (45°, 90 °) (326°, 89 °)

MACI P 350 MSVF P 104 WVT P 275 PAB SH 418 HWUT SH 1223
(47°, 79 °) (244°, 97 °) (346°, 69 °) (45°, 94 °) (331°, 81 °)

SACV P 392 NIUE P 229 WCI P 254 MACI SH 526 BW06 SH 1340
(52°, 65 °) (249°, 86 °) (347°, 70 °) (47°, 79 °) (332°, 81 °)

KOWA P 433 RAR P 238 DWPF P 284 RCBR SH 814 RSSD SH 1351
(65°, 79 °) (252°, 77 °) (350°, 60 °) (60°, 42 °) (337°, 80 °)

DBIC P 400 AFI P 160 KAPO P 242 TSUM SH 557 CRIN SH 1694
(71°, 74 ) (252°, 91 °) (352°, 81 °) (106°, 79 °) (339°, 46 °)

SHEL P 444 PTCN P 438 SADO P 266 LSZ SH 408 JTS SH 1726
(91°, 61 °) (261°, 51 °) (354°, 76 °) (108°, 90 °) (340°, 43 °)

TSUM P 335 XMAS P 241 SSPA P 275 LBTB SH 653 ULM SH 1213
(106°, 79 °) (273°, 87 °) (355°, 72 °) (115°, 83 °) (344°, 84 °)

LSZ P 230 POHA P 131 GAC P 302 BOSA SH 847 BCIP SH 1734
(108°, 90 °) (290°, 95 °) (357°, 76 °) (118°, 80 °) (347°, 41 °)

SUR P ° 371 KIP P 116 MNTQ P 262 SUR SH 673 MTDJ SH 1292
(119°, 75 °) (290°, 97 °) (358°, 76 °) (119°, 75 °) (352°, 49 °)

HOPE P 481 SLBS P 334 QSPA SH 870 KAPO SH 1073

(142°, 33 °) (321°, 66 °) (180°, 58 °) (352°, 81 °)
0 30 60 90 120
ORCD P 314 PFO P 289
Time (s) CASY SH 781 GTBY SH 1448
(156°, 34 °) (323°, 77 °) (180°, 82 °) (355°, 51 °)

ESPZ P 416 COR P 202 SBA SH 1119 LBNH SH 1237

(168°, 33 °) (326°, 89 °) (191°, 65 °) (359°, 75 °)

Figure 5
Observed (black traces) and synthetic (red traces) broadband P and SH ground displacements at global stations, at indicated azimuths and
epicentral distances. The peak-to-peak amplitude of the data (in microns) is indicated for each trace on the right. The synthetics are for the
finite-fault model shown in Fig. 4d

indicate a down-dip region of *5 m slip extending
below the coastline 30–60 km north of the
hypocenter.
2.4. Source Spectrum, Radiated Energy and Stress
Drop Estimation
The moment rate function for the model in Fig. 4d
(Vr = 2 km/s, dip = 22°) is shown in Fig. 6a, and
there is a 95.5-s total duration, with a seismic
moment of 2.67 9 1021 Nm (MW 8.2). The centroid
time is 50.5 s, several seconds larger than the gCMT
centroid time, which suggests that the last 10 s of the
source model may be model noise. The overall shape
of the moment rate function, which has very little
dependence on the specific finite-fault inversion
model, is very smooth and close to a truncated
Gaussian in symmetry. This contributes to scalloping
of the smooth source amplitude spectrum, with a deep
notch near 0.03 Hz as shown in Fig. 6b. This source
spectrum is constructed from the spectrum of the
moment rate function for frequencies below 0.05 Hz
and from log-averaging of far-field P wave displace-
ment spectra corrected for radiation pattern,
geometric spreading, and attenuation for frequencies
above 0.05 Hz. A reference x-2 spectrum with
3 MPa stress parameter and the same seismic
moment is shown by the dashed line.

Reprinted from the journal 18

Vol. 173, (2016) Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel

(a) (c)
M = 2.67 x 10 SH (42)
8 MW = 8.22
M (x 10 Nm/s)
Tc = 50.5 s, Td = 95.5 s
6
4
2
.

0
0 20 40 60 80 100 120
Time (s)

(b) φ = 5.0°, δ = 22.0°, λ = 95.8°

10

P (60)
Moment (N−m)

10

10

Er = 2.93 x 10
10
Er /M0 = 1.10 x 10

10 10 10 10
Frequency (Hz)
Figure 6
a Moment rate function for the 16 September 2015 Illapel earthquake from the teleseismic body wave finite-fault inversion in Fig. 4d. The
seismic moment, MW, centroid time Tc, and total duration, Td estimated from this model are shown. Note the overall Gaussian shape of the
source function. b The average source spectrum for the 2015 earthquake constructed using the moment rate function spectrum in (a) for
frequencies below 0.05 Hz (black line) and log-average spectra of teleseismic P wave ground displacements corrected for geometric
spreading, radiation pattern, and attenuation for frequencies above 0.05 Hz (red line). The total radiated energy integrated to 1 Hz, Er, and the
seismic moment-scaled values are indicated. Note the deep scallop in the source spectrum near 0.03 Hz; this is a result of the smooth truncated
Gaussian shape of the moment rate function. The SH and P wave radiation patterns and data sampling for the average focal mechanism from
the finite-fault inversion are shown in (c)

The far-field radiated energy for the source was estimate posted by IRIS is 3.2 9 1016 J (http://ds.
estimated from teleseismic P wave ground velocities iris.edu/spud/eqenergy/10095476) based on the
following the basic procedure of VENKATARAMAN and method of CONVERS and NEWMAN (2011).
KANAMORI (2004), adding relative contribution from For the slip models in Fig. 4 we made two
low-frequency energy based on the moment-rate calculations of the stress drop. In a simplified proce-
spectrum following the procedure of YE et al. dure, we remove the subfaults with inverted seismic
(2013a). For the case shown in Fig. 6, this gives a moment less than 15 % of the peak subfault seismic
radiated energy estimate up to 1 Hz of moment and then use the total area and average slip of
Er = 2.9 9 1016 J, and a moment-scaled value Er/ the remaining subfaults (outlined in Fig. 4) in a
M0 = 1.1 9 10-5. The various models yield a range uniform slip circular crack model, finding estimates
of radiated energy estimates because of the differ- of Dr0.15 = 1.7–3.4 MPa. To better account for the
ences in depth of the slip distributions, spanning spatially varying slip effects we also calculate the
values from 2.2 9 1016 to 3.15 9 1016 J, with Er/ shear stress near the center of each subfault (Fig. 7
M0 = 0.8–1.1 9 10-5. The broadband energy shows the result for the model in Fig. 4d) and integrate

19 Reprinted from the journal


−60
Distance along dip (km)
−30
0 28.0
30
−120 −90 −60 −30 0 30
Distance along Strike (km)
0 1 2 3 4
Shear Stress (MPa)
5 6
Figure 7
The distribution of stress change associated with the finite-fault slip
model for the 16 September 2015 Illapel earthquake shown in
Fig. 4d, where the stress values are calculated near the center of
each subfault. The black arrows indicate the direction of the shear
stress, with the color-coding indicating the value for each subfault.
The slip-weighted integral of stresses is used to calculate the stress
drop DrE, whereas a simplified procedure that trims off subfaults
with moments less than 15 % of the peak subfault moment and uses
the remaining subfaults to estimate the rupture area and average
slip for a circular stress drop estimate gives Dr0.15
the slip-weighted stress change to find estimates of
DrE = 2.6–4.0 MPa following the procedure of NODA
et al. (2013). The corresponding estimates of radiation
efficiency gr = 0.17 (for the 16° dip models) to 0.47
(for the 22° dip model with Vr = 2.0 km/s) are
somewhat low values due to the relatively low
moment-scaled radiated energy (YE et al. 2013b).
3. Discussion and Conclusion
The teleseismic parameters estimated here are not
unique, both due to the limitations of seismic wave
coverage and dependence on assumed parameters such
as the dip. Joint analysis of seismic, geodetic, and
tsunami data will soon establish more definitive values
for the rupture process. However, it is valuable to have
rapidly obtained first-order seismic solutions for
applications such as tsunami warning as well as for
comparison with additional non-unique analyses of
geodetic and tsunami data. The W-phase method
applied to regional data provided a stable mechanism,
seismic moment, and depth obtainable within
5–10 min after the event, which was valuable for
tsunami warning. Other seismic source attributes noted
here were obtained within a few hours of the event.
Reprinted from the journal
L. Ye et al. Pure Appl. Geophys.
5.5
The models with large slip near the trench are
representative of all models we found with shallower
dip values. This attribute is shared by the USGS-
Depth (km)
16.8
NEIC finite-fault solution for a dip of 19° (http://
earthquake.usgs.gov/earthquakes/eventpage/us20003
k7a-scientific_finitefault), and with some other on-
39.2 line posted models (e.g., http://www.geol.tsukuba.ac.
jp/*yagi-y/EQ/20150917/index.html). The along-
60 90 120 150 strike placement of slip varies with the assumed
Δσ0.15 = 1.65 MPa
rupture velocity, but in general such models place the
ΔσE = 2.63 MPa slip up-dip of most aftershocks. The models with
steeper dip do not extend as far off-shore, but tend to
put the slip deeper on the fault as well, closer to the
coast, in closer proximity to most aftershocks
(Fig. 1b) and near the seaward region of short-period
radiation imaged by back-projection (Fig. 2c). This is
consistent with some other on-line posted models
(e.g., http://www.earthobservatory.sg/news/september-
16-2015-chile-earthquake). Any one-dimensional
structural model does not account for the precise
depth below seafloor and intersection with the trench.
The moment rate function is relatively robust for
variations of rupture velocity, although the moment
varies with dip. The relatively low rupture velocities
of 1.5–2.0 km/s indicated by the high-frequency
back-projections limit the along-strike expansion of
the rupture. The lack of short-period bursts of energy
from the large-slip regions near the trench indicated
by the shallow dip models is similar to what has been
observed for numerous large events (LAY et al. 2012).
The steeper dip models do not have as strong of a
discrepancy between the large-slip region and the
back-projections, although the strongest feature in the
back-projections is still deeper than the main slip
patch. The moment-scaled radiated energy value is
higher than typical of tsunami earthquakes (e.g.,
YE et al. 2013b) and this event does not have the
distinctive long centroid time shift of a tsunami
earthquake. The report of a localized 4.7 m tsunami
near Coquimbo suggests concentrated slip offshore
northward of the hypocenter, compatible with the
steeper dipping models we considered. The overall
tsunami observations do not indicate the exception-
ally large tsunami expected for a near-trench tsunami
earthquake.
The peak slip estimates of 8–12 m indicated by
our teleseismic finite-fault inversions raise questions
20
Another random document with
no related content on Scribd:
royalists, they were able to carry a resolution rendering members of
the existing legislature ineligible for election to the next, and thus
driving their most active opponents, for a time at any rate, out of
power. The revision of the Constitution told slightly against them, but
it came to very little, and all its worst faults were retained. When the
Constituent Assembly separated at the end of September,
Robespierre and Pétion, not Lafayette or Barnave, were its heroes
with the populace of Paris. And in the months which followed, the
power of their party increased. In spite of the motion which the
majority of the Assembly had passed just before its dissolution,
forbidding the affiliation of popular clubs and their interference in the
general election, the Jacobin clubs rapidly multiplied, and threw all
their energies into the electoral contest. The abolition of the property
qualification for deputies had already been secured by the
persistence of Robespierre. The retirement of Lafayette from the
chief command of the National Guard, and the abolition of that post
as a permanent office, considerably weakened the Constitutional
party. The growing sense of weariness with politics, and the desire to
rest from agitation felt by the bulk of the people, began to show itself
more distinctly. The renewal of the whole of the legislature and of
one-half of all the local officials, afforded an opportunity for many
moderate and experienced men to retire, and for more pushing and
ambitious politicians to fill their places. The number of elections, and
the fact that they were held so near together, prevented many voters
from recording their votes. The necessity of taking the oath to
observe the new ecclesiastical system disfranchised a large number
of scrupulous Catholics. The intimidation practised by the Jacobins
against all reactionary voters, of which there are clear examples but
the effects of which it is difficult to estimate, must have kept many
quiet people away. All over France the proportion of voters who
came forward to vote was very small. The result showed, in the new
Assembly, a considerable increase in the advanced party, and many
new-comers who hurried at once to join the Jacobin Club. But it
showed also that the majority of voters loyally adhered to the new
Constitution, were fully prepared to give it a trial, and were well
represented even in Paris itself.
Unhappily, however, for France, the majority never found the time to
rally. From the end of 1791, the shadow of war began to darken the
political horizon. At the critical moment, when the nation had to
choose between the majority, which wished to consolidate the new
system, and the minority, which wanted to destroy it for something
else, the alarm of invasion redoubled the panic and disorder,
paralysed any possible reaction, and threw Frenchmen off their
balance again. The war with Europe meant a struggle both for
freedom and for national existence. In the tumult of that struggle all
other considerations were flung to the winds. And the men who could
best save the Revolution and maintain the honour of the country
became the heroes and the tyrants of France.

FOOTNOTES:
[7] The 144 representatives of the Sections formed the General
Council of the Commune; 48 of these formed the municipal body;
and 16 of these were the actual Administrators, distributed among
five departments, of 'subsistence,' 'police,' 'finance,' 'public
establishments,' and 'public works.' See Mortimer-Ternaux's
Histoire de la Terreur, vol. i. Appendix III, where the whole subject
is thoroughly discussed.

CHAPTER VII.
The Influence of the War upon the Revolution.
It is a mistake to imagine that the European Powers attacked the
Revolution in France. It was the Revolution which attacked them.
The diplomatists of the eighteenth century viewed at first with cynical
indifference the meeting of the States-General at Versailles. They
were naturally blind to its significance. It did not occur to them, until
some years had passed, that the outbreak in France was anything
more than a temporary political disturbance, which by weakening a
redoubtable rival, would redound to the advantage of the other
Powers. The new cosmopolitan spirit, the new idea of enthusiasm for
humanity, which, together with the influence of Rousseau, were
beginning to be deeply felt in Europe, did not affect the diplomatic
mind. Many a sharp lesson was needed to convince the cabinets of
Europe that events in France were not the result of any ordinary
political commotion, but part of a movement as far-reaching as it was
profound, abstract in its aims and maxims, universal in its scope,
irresistible in its advance, and inspired by the propagandist
enthusiasm which in earlier ages had produced the great Crusades
and the religious wars.
Accordingly, the attitude of the European Powers was at first one of
complacent egotism. They viewed the disorders in France with the
suave but moderate concern which nations, if not men, display
towards the misfortunes of their friends. Moreover, in 1789, most of
the Powers were occupied with matters of far greater urgency at
home. On the throne of Russia sat the most brilliant sovereign of the
age, the sovereign to whom Diderot attributed the 'soul of Brutus'
and 'the form of Cleopatra,' steeped in all the vices with which the
imagination of her enemies credited Theodora, excelling in all the
accomplishments with which the enthusiasm of her subjects clothed
Elizabeth, unsurpassed in ambition and ability, equally unwearied in
war and diplomacy, in literature and love, a singularly sagacious
ruler, an extraordinary woman and a most successful queen. By the
side of Catherine there reigned in the North another able and
attractive prince, Gustavus III of Sweden, who combined with
restless vanity and philosophic maxims a real capacity for
government and the spirit of 'a Saladin in quest of a crusade.' In
Austria, the Emperor Joseph, the madcap crowned philosopher of
Europe, had thrown his dominions into confusion by his reforms, and
threatened to convulse the East with his ambition. In Prussia,
Frederick William II, with his fine aspirations and his fantastic
failings, permitted favourites and charlatans to help him in
maintaining the great traditions of his predecessor. In Spain, the
declining Bourbon dynasty vainly looked to France to save it from
losing its once proud position in the world. In England, Pitt, at the
height of his power, was ready to offer France an alliance which she
was too suspicious to accept from Chatham's son, and while
anxiously watching the troubles in Turkey, and successfully asserting
his influence in Holland, fixed his hopes on the wide humanitarian
projects which he saw opening out before him, to which his genius
and his inclination called him, and which he hoped to consummate in
peace.
The two points which occupied the attention of Europe in 1789 were
the condition of Poland and the troubles in the East. The ambitious
designs of Catherine and the assistance lent to them by Joseph
threatened the existence of the Turkish Empire, irritated the Prussian
Court, and awakened English apprehensions, always sensitive about
the safety of Stamboul. Poland, the battle-field of cynical diplomacy,
torn by long dissensions and ruined by a miserable constitution, was
vainly endeavouring, under the jealous eyes of her great neighbours,
to avert the doom impending, and to reassert her ancient claim to a
place among the nations of the world. But Russia had long since
determined that Poland must be a vassal State to her or cease to be
a State at all, while Prussia, driven to face a hard necessity, realised
that a strong Poland and a strong Prussia could not exist together,
and that if Poland ever rose again to power, Prussia must bid good-
bye to unity and greatness. These two questions to the States
involved seemed to be of far more moment than any political reform
in France, and engrossed the diplomatists of Europe until the
summer of 1791.
In February, 1790, a new influence was introduced into European
politics by the death of the Emperor Joseph and the accession of his
brother, Leopold II. Leopold was a man of remarkable ability, no
enthusiast and no dreamer, thoroughly versed in the selfish traditions
of Austrian policy and in some of the subtleties of Italian statecraft,
discerning, temperate, resolute and clear-headed, quietly determined
to have his own way, and generally skilful enough to secure it.
Leopold found his new dominions in a state of the utmost confusion,
with war and rebellion threatening him on every side. He speedily set
about restoring order. He repealed the unpopular decrees of Joseph.
He conciliated or repressed his discontented subjects. He gradually
re-established the authority of the Crown. He had no sympathy with
Joseph's Eastern schemes, and he dreaded the advance of Russia.
He was resolved that Prussia should no longer assume the right to
intervene decisively in European politics, but should return to her old
position as a dependent ally of Austria. He hoped by steady and
persistent diplomacy to secure a reform of the Polish constitution,
and to build up in Poland once again a State in close alliance with
himself, strong enough to hold in check Russia and Prussia alike. He
was not blind to the gravity of the events happening in France, and
his sympathy for his sister's misfortunes was sincere. But he saw
more clearly than most men the very great difficulties involved in any
intervention in French affairs, and although he showed his feelings
and was ready to discuss proposals for bringing the influence of the
Powers to bear upon French politics, he was far from wishing to
commit himself to any irremediable breach with France.
Accordingly, the first eighteen months of Leopold's reign were
occupied with his own immediate interests, and at the end of that
time his success was marked. Catherine's vast schemes in Turkey
had been checked. War had been averted. Poland had been
strengthened by internal changes. Prussia had been conciliated and
out-manœuvred, and her influence had been impaired. At last, at the
end of August, 1791, the Emperor was free to face the French
problem, and he set out for the Castle of Pillnitz to meet the King of
Prussia and the Emigrant leaders at the Saxon Elector's Court.
For some time past the restlessness of the French Emigrants had
been causing great perplexity in Europe. Received with open arms
by the ecclesiastical princes of the Rhine, by the Electors of
Mayence and Trèves, they proceeded to agitate busily for their own
restoration. They brought with them into banishment all the worst
characteristics of their class. They treated their hosts with cool
impertinence, indulged in the most complacent forecasts, and
exhibited that profound contempt for anything except their own
advantage, which had made them so justly detested in France. In the
view of the Emigrants their cause was the cause of feudal Europe.
Just as the revolutionists regarded all nobles as enemies, so the
Emigrants regarded all princes and aristocrats as companions in
arms. Deeply and serenely selfish, they would hear of nothing but
their own reinstatement in the full privileges of the Ancien Régime;
and all compromise with constitutional freedom, all moderation in
counsel or in language, all respect for the wishes of the King was
treated with disdain. Rarely has any class of men displayed in a
more conspicuous manner its lack of patriotism, sagacity and
temper, and its utter indifference to any interest but its own.
The object of the Emigrants was to bring pressure to bear at the
European Courts, with the view of inducing the Powers to intervene
actively in their behalf. The Comte d'Artois and his Prime Minister,
Calonne, flitted from capital to capital, to plead the cause of the
oppressed aristocracy of France. At Vienna they exposed
themselves to unmistakable rebuffs, but in other quarters they met
with kinder treatment. Prussia assured them of her sympathy, but
regretted that she could take no action without the Emperor's
concurrence. Sardinia and Spain were friendly, and talked frequently
of war. Catherine of Russia, who cordially hated the doctrines of the
Revolution, but who under-rated the importance of a movement
which was too far off to touch her people, exchanged delightful
compliments with the descendants of Henri Quatre, and protested
her enthusiasm for their cause. But the prudent Empress showed no
inclination to risk the life of a single Cossack in propping up the
Bourbon throne. Gustavus of Sweden, appearing at Aix-la-Chapelle,
enlisted with ardour in the Emigrants' service. He proposed to
transport a joint army of Swedes and Russians to the coast of
Normandy, and marching up the Seine to Paris, to assert the rights
of kings. But when Gustavus applied to Catherine to assist him in
carrying out his crusade, he discovered, to his bitter disappointment,
that no assistance was to be obtained.
Nevertheless the lofty tone of the French princes did not alter. After
his escape from France, in June, 1790, the Comte de Provence
established his Court at Coblentz, where he was joined by his
brother the Comte d'Artois, and where, on the plea that Louis was a
prisoner, he claimed the title of Regent, and assumed the authority of
King. The Court of the two French princes at Coblentz represented
faithfully the faults and follies of the Emigrant party. But a more
satisfactory spectacle was offered by the camp at Worms, where
Condé was bravely trying to organise an army to fight against the
Revolution in France. To Condé's standard flocked the more patriotic
Emigrants, who disliked the idea of foreign intervention and hoped to
recover their position for themselves. Condé had no difficulty in
finding officers, but privates and troopers were more difficult to raise.
Nevertheless he persevered in his enterprise, and gradually
collected a considerable force[8]. In the winter of 1791-92, a plot for
the capture of Strasbourg was discovered, which greatly alarmed the
French Assembly. But the German Princes in the neighbourhood
looked with disfavour on the Emigrant army. It caused confusion in
their dominions, and it drew down on them the hostility of the French
Government. The Emperor joined them in protesting against it. In
February, 1792, Condé's army was compelled to abandon its camp
at Worms, and to retire further into Germany.
The Emperor was well aware of the reckless selfishness of the
Emigrant princes. He had as little sympathy with them as his sister.
He did not intend to listen to their demands. If he interfered in France
at all, it would only be in a cautious and tentative manner, and in
order to save Marie Antoinette and her husband. Certainly he would
not undertake a war for the restoration of the Ancien Régime. The
real inclinations of Leopold, with which Marie Antoinette generally
concurred, pointed towards the summoning of a European
Conference, to bring pressure to bear on the French Government, to
strengthen the hands of the moderate parties in Paris, and to prevent
any outrage on the King and Queen. Accordingly, the interviews at
Pillnitz came to nothing. The Emperor and the King of Prussia did
indeed enter into a treaty, binding them in certain events to make
war together upon France. But the conditions of that alliance, which
involved the joint action of all the European Powers, and in particular
the co-operation of England which was known to be opposed to any
form of intervention, rendered the treaty an empty form. Leopold
himself assured his Chancellor that the Conference of Pillnitz had
bound him to nothing, and the whole proceeding has been rightly
described as one of the 'august comedies' of history. As time went
on, the Emperor's determination to avoid war deepened. When, in
September 1791, Louis accepted the new Constitution, Leopold took
advantage of that event to make his position unmistakably clear. On
the 1st November, he addressed a circular letter to the Powers,
pointing out to them that Louis' action rendered any further
interference impossible, and in spite of the angry protests of the
Emigrants, the danger of war seemed to have disappeared.
Such was the attitude of Europe when the Legislative Assembly met.
Numerically, the majority of the new Assembly were favourable to
the Constitution, and honestly wished to maintain the throne. Outside
the House they were supported by Lafayette and Barnave, two
strong, representative figures, although Lafayette could not be
depended on to act consistently with any colleagues. But in the
House they had no leaders of great influence, and no sense of party
discipline. Their policy was negative and undecided, and they could
not always be relied on to vote together. Opposite to them there sat,
on the left of the Assembly, a body of deputies less strong in
numbers, but far more able and united, full of eloquence and
enthusiasm, eager, brilliant, reckless and impetuous, with a clear
policy and bold, ambitious views. It consisted of men who despised
the Constitution as an unsatisfactory makeshift, who had no
reverence for the throne, who wished for no compromise with the old
order, but who, conceiving themselves inspired with the spirit of the
heroes of Greece and Rome, were resolved to sweep away kings
and tyrants, to undertake a crusade in the name of liberty, and to
establish triumphantly their republican ideal in France. To this body
the Jacobin deputies attached themselves, because for the present
their aims were the same. But the greater part of it was composed of
men who, though in temporary alliance with the Jacobins and ardent
recruits of the Jacobin Club, were soon destined to become their
rivals, and who are distinguished as the Girondist party, because
their chief leaders came from the department of the Gironde.
The Republican minority in the Legislative Assembly, supported by
the Jacobins and directed by the Girondist leaders, easily out-
manœuvred their Constitutional opponents. Their object was to
discredit the Monarchy and so to prepare the way for a Republic.
From the first, the Girondists were strong partisans of war, partly
because their reckless patriotism wished to make all 'tyrants tremble
on their thrones of clay,' but also because they had a just conviction
that a policy of war would play into their hands, would render the
king's position desperate, and would promote confusion in which
they would win. Vergniaud, Guadet, Gensonné and others clothed
the Girondist sentiments in language which, if sometimes bombastic,
was sometimes superb. A House full of inexperienced theorists, all
impulsive and nearly all young, echoed the fine enthusiasm which
the Girondist orators expressed. Brissot, a journalist of many strange
experiences, but with a very shallow knowledge of affairs, took the
lead of the advanced party, and guided their views upon diplomatic
questions. Madame Roland appealed to their emotions with
generous rhetoric and hospitality. Sieyès, moving mysteriously in the
background, and pulling the strings of innumerable intrigues, drew
up their plans and assisted at their counsels. Under these
circumstances the Girondist minority, with its vigour, its ambition, and
its somewhat unscrupulous designs, seized the control of events,
and committed the halting, uncertain majority to measures from
which they could not afterwards retreat.
The autumn and winter of 1791-92 were occupied by schemes and
counter-schemes of every kind. The whole atmosphere was charged
with intrigue. The records of each week were full of evidences of
disorder, to which those in power seemed to pay little heed. The
condition of Europe was already alarming, and the Girondists, in
search of their 'second and greater Revolution,' resolved to turn it to
account. The Assembly knew nothing of diplomacy, nothing of
Leopold's peaceful views, nothing of the Queen's reluctance to face
the risks of war. They saw the sovereigns of Europe combining
together. They saw the Emigrants gathering outside the frontier, and
the majority of the old priesthood stirring up discontent within. They
believed, or the Republicans among them believed, that the King
was urging on the Powers to war, and that the Queen presided over
a secret 'Austrian Committee,' which, meeting in the Tuileries,
conspired against the liberties of France. Accordingly, they
proceeded to denounce these dangers and to demand energetic
measures against them. In October and November, decrees were
passed, calling on Monsieur and the Emigrants to return,
commanding the refractory priests to accept the Constitution, and
imposing heavy penalties for disobedience. The decree directed
against his own brother Louis confirmed, but the other two decrees
he vetoed; and of course the application of the veto still further
increased his unpopularity in Paris.
As the winter went on, the gravity of the situation deepened. Bailly
retired from the mayoralty of Paris, and Pétion, the nominee of the
Republican party, was elected in his place over the head of
Lafayette, by a large majority on a very small poll. Montmorin, the
Minister of Foreign Affairs, resigned, and was succeeded by De
Lessart, a man of indifferent capacity, much under the influence of
Barnave and the Lameths. Other ministerial changes followed.
Bertrand de Molleville, a strong supporter of the Crown, was
appointed Minister of Marine; and early in December, Narbonne, a
brilliant young man of fashion and ability, lifted into power by the aid
of Lafayette and by the intrigues of Madame de Staël, became for a
short time Minister of War. But these changes did nothing to
strengthen the Government. While De Lessart and the Lameths
dreaded war, and hoped that the Powers would exert some pressure
in favour of the moderate party in France, Narbonne threw himself
heartily into the military preparations. He believed that war would
strengthen the Crown, and he determined to make a bid for popular
favour on the King's behalf by identifying Louis with the patriotic
feeling. Meanwhile Lafayette, with characteristic indiscretion, blind to
the warnings of the wiser Constitutional leaders, possessed by his
favourite idea of carrying American freedom over Europe, and
tempted perhaps by the great prospect which war would open to his
own ambition, supported Narbonne in his warlike designs, and
steadily encouraged the war-party in the Assembly.
To the surprise of the Girondists, the Jacobin leaders took a totally
different line. Fearing that the declaration of war would increase the
power of the Government and would strengthen the hands of the
hated Lafayette, the Jacobins began to sever themselves from the
Gironde, and to oppose the idea of a campaign. No doubt, their
opposition was partly due to the fact that on principle many of them
were averse to war, although in the days of the Jacobin triumph, this
principle, like others, was to be cast to the winds. But in so far as
their opposition was due to tactical and party reasons, it shows a
strange lack of political discernment, for of all parties then existing
the Jacobins were the most certain to profit by the outbreak of
hostilities. Brissot and the Girondists saw this clearly, and vainly
endeavoured to convince their allies. From that time forward the
rivalry between Robespierre and Brissot became bitter and acute.
But the Jacobin politicians allowed their fear of the Executive to carry
them away. Anything which made rulers powerful must, they thought,
be dangerous to freedom. Danton, Robespierre, Marat and Billaud-
Varennes, all adopted the same language, and the Jacobin Club
protested loudly against the demand for war.
But the Girondists carried the day. Lafayette's manœuvres, their own
enthusiasm, and the militant temper of the nation, all helped their
designs. Narbonne's schemes indeed collapsed, and early in March,
1792, he was dismissed from a Ministry with most of which he
disagreed. But the war-party revenged Narbonne's dismissal by
driving his colleagues from office. Louis, yielding to the storm, and
endeavouring once again, as he had sometimes endeavoured
before, to identify himself with the Assembly, selected a new Ministry
from the popular party. Roland was appointed Minister of the Interior,
and Dumouriez took the portfolio of Foreign Affairs. Dumouriez, who
owed his appointment to the Girondists, for whom, nevertheless, he
had a rich contempt, was the only person of note in the Government,
but he possessed ability enough to compensate for all the
deficiencies of his colleagues. Lax in morality and principle, he was a
man of infinite resource, bold, ambitious and consummately adroit.
He welcomed the idea of a conflict with Austria. He hoped, as
Narbonne had done, to secure the neutrality or alliance of England,
and, if possible, the friendship of Prussia, but he was prepared to
take the chances of a struggle with the rest of Europe. The
appointment of the new Ministry gave the Girondists the command of
the political situation, and from that moment France drifted rapidly
into war.
Events abroad made a rupture easy. In spite of the provocations
offered to him by the French Assembly, Leopold had clung steadily
to peace. His sagacity saw that the one chance for the Monarchy in
France lay in the desire of the Constitutional party to re-establish
order. He was determined to strengthen their hands, if he could, and
for that purpose to limit the interference of the Powers to joint
diplomatic pressure in their behalf. But early in March, 1792, Leopold
suddenly died. His heir Francis, unrestrained by his father's tact and
moderation, assumed a different tone and showed less patience.
The chances of any effective pressure from the Powers declined, as
the prospect of war rose on the horizon. Francis' language was
sufficiently sharp to give the Assembly the pretext which it longed for,
and on the 20th April, Louis, amid general enthusiasm, came down
to the Assembly and declared war against Austria. The effects of that
momentous step no comment can exaggerate. It ruined the best
hopes of the Revolution, and prepared the way for a military
despotism in the future. All who hesitated, all who felt that mistakes
had been committed but who still hoped that they might be repaired,
all who believed that the Revolution might yet vindicate itself by
combining liberty with order, saw themselves forced to choose, no
longer between order and disorder, but between the old system and
the new, between the ancient Monarchy and freedom, between the
cause of their country and the cause of the invaders. Had there been
no war with Europe, the astonishing episode of the Jacobin triumph,
the worst excesses of the later Revolution, and all the crimes and
glories of the Terror, could never have taken place in France.
It should be clearly understood that, even after the declaration of
war, the friends of the Monarchy, who wished the Revolution to
pause, were in a great majority both in Paris and in France. But they
were disorganised and often lukewarm, divided into numberless
different groups, jealous and distrustful of one another, largely
governed by personal motives, with no clear policy before them,
incapable of acting loyally together, and without the ability to act
wisely, even if they could unite. Barnave and the Lameths distrusted
Lafayette. Lafayette distrusted Dumouriez. The stronger royalists
distrusted Barnave. The King distrusted all alike. On the other side
was a smaller but more active party, full of enthusiasm and audacity,
not, it is true, without enmities and divisions of its own, but still better
organised than its opponents, prepared to embark on a policy of
danger, and to hope that the future would turn to its account. In that
situation of affairs the war broke out, and its effects were soon
apparent. It rendered Louis' dethronement almost certain. It drew a
sharp line between those who were on the side of France and of the
Revolution, and those who were on the side of the Emigrants and
invaders, a line which placed the King upon one side and the vast
majority of his subjects on the other. It rallied all patriots to the party
which undertook the national defence. It made the Jacobins, whom
the enemy denounced, the heroes of the popular resistance. It
forced into helpless inaction all those who wanted order and a king,
but who could not lift voice or hand for Louis, if by doing so they
weakened the unity of France. It rendered possible, though it did not
necessitate, the Terror, for it converted all opponents of the
Revolution into traitors. It led at once to national peril, and through
peril to panic and confusion. In the confusion the elements of
disorder, already rife in France and taught impunity by the
experience of the past two years, rose uncontrolled in insurrection,
and patriotism became identified with crime.
The movement of events was rapid. At the end of April, the
campaign opened with an attempted invasion of Belgium. But the
French troops were disorganised; their commanders were timid and
incapable; two French divisions were shamefully defeated, and the
general of one was murdered by his men. 'You marched out like
madmen,' wrote Dumouriez bitterly, 'and you returned like fools.' The
bad news from the front intensified the excitement in Paris. Another
Ministerial crisis resulted in the appointment of Servan, a stern
Republican, as Minister of War. The attacks upon the Queen
redoubled. The lawlessness of the politicians of the streets
increased. The Girondists, determined to weaken the Monarchy,
abolished the King's constitutional guard, voted the banishment of all
refractory priests, and decreed that a camp of twenty thousand men
from the departments should be formed in the immediate
neighbourhood of the capital. Since the winter, the Parisian mobs
had been armed with pikes; and it seems that the Girondists,
knowing that the influence of the Jacobin leaders with the populace
of Paris was greater than their own, desired to have at hand a strong
force of ardent revolutionists, distinct from the Parisians, on whom
they could rely. The King, however, disliking these proposals, and
wearied by the studied rudeness of his Ministers, refused to sanction
the formation of the camp and the persecution of the refractory
priests. Then Madame Roland, in her husband's name, attacked him
in a foolish and impertinent letter, and Louis, roused to unusual
irritation, dismissed his Girondist advisers on the 12th June. For a
moment Dumouriez remained at the head of affairs; but finding that
he could not induce Louis to accept his views, he too resigned a few
days later, and accepted a command in the army. Lafayette took
advantage of the occasion to make a violent attack upon Dumouriez,
thus converting into an enemy a man whom he might have found an
invaluable ally. Louis fell back on a new Ministry of personal friends
of Lafayette, and the General wrote to the Assembly denouncing and
threatening the Jacobin party. Thereupon the mob forces of Paris,
equally alarmed by the dismissal of the Girondists and by the tone of
Lafayette, broke into insurrection and invaded the Tuileries on the
20th June.
The insurrection of the 20th June, which had been for some time
preparing, was not the work of the Republican leaders. The
Girondists held aloof, and Danton and Robespierre discouraged the
proposal. It was entirely the action of the subalterns of the party, led
by Santerre. Almost to the last, the responsible men held back. Even
on the 10th August, the Girondist leaders, who had been working for
months to upset the throne, hesitated, when the victory was within
their grasp. They had grown afraid of their Jacobin allies, while the
Jacobins knew the smallness of their own forces, and still feared the
strength of their opponents. The abortive riot of the 20th of June was
followed by a brief reaction in the King's favour. Lafayette came up to
Paris, denounced the rioters at the bar of the Assembly, proposed to
shut the Jacobin club, and offered to carry the King out of Paris. The
National Guards and the Parisian bourgeois, shocked by the insult
offered to the Sovereign, showed themselves ready to rally round
Lafayette. One of the new ministers, Terrier de Monciel, was a man
of considerable energy and insight. He urged the King to place
himself in the hands of the Constitutional party, and with their help to
escape from Paris and appeal to France. But Louis, even in his
desperate situation, could not be prevailed on to act cordially with
Lafayette. He preferred to trust to the chapter of accidents and to
wait for the Allies to deliver him. The General's offers were coldly
received. The favourable moment was allowed to pass. Lafayette,
naturally offended, and always incapable of decisive action, returned
humiliated to his army. Monciel's schemes were rejected, and, early
in July, he and his colleagues resigned. The Court had deliberately
thrown away its last chance of safety.
All through the month of July the agitation in Paris increased. On the
11th, the National Assembly declared that the country was in danger,
and issued a stirring appeal for volunteers. The Republicans began
to rally again, and the arrival of the Fédérés from the departments, to
celebrate the festival of the fall of the Bastille, although partly
checked by the vigorous action of Monciel, supplied them with the
force which they required. The troops of the line remaining in the
capital were ordered to the front. The leaders of the insurrection of
the 20th June, acting with the Fédérés from the provinces, and
encouraged this time by the Jacobin leaders, set to work to organise
a rising in the revolutionary Sections of Paris. The denunciations of
the King and of Lafayette, and the intimidation of the deputies in the
streets redoubled. The reactionary party heightened the excitement
by prophesying the speedy vengeance of the invaders. The allied
forces at last began to show signs of activity, and at the end of July,
the Duke of Brunswick, their commander-in-chief, issued a manifesto
to the French people.
The idea of a manifesto had originated with Louis himself. Before the
end of May, acting on the advice of Malouet and Montmorin, he had
sent Mallet du Pan with a confidential message to the Allied
Sovereigns at Frankfort. In this communication he entreated the
Powers to adhere to Leopold's policy, to make it plain that their
object was not to dismember France or to restore the proscribed
classes, but only to set the King free, so that he might suppress the
Jacobins, and readjust the Constitution in the interests of order and
liberty combined. Louis' objects were not unworthy, but the policy by
which he sought to achieve them was hopelessly unsound. Even had
the Allies taken his advice, no arguments could have made the
invasion palatable to Frenchmen. And, once war was declared, there
was little chance that his advice would be heeded. Since the
outbreak of hostilities the influence of the Emigrants had increased
abroad. They paid no attention to the warnings of Louis. They
indignantly discarded the moderate language suggested to them by
Mallet du Pan, and imparted their own spirit to the Duke of
Brunswick's manifesto. The result was that the manifesto, with its
violent and irrational menaces, caused the wildest indignation in
France, roused the whole people to protest against it, and
immediately facilitated the deposition of the King.
At last, early in August, the crisis came. The manifesto of the Allies,
the arrival of a body of zealous Republican Fédérés from Marseilles,
and the final breach between the King and the Gironde, precipitated
the insurrection. The municipal authorities distributed cartridges
freely in the Faubourg St. Antoine, while they refused ammunition to
the National Guard. The insurrectionary leaders in the Sections
completed their organisation, and arranged matters with the officials
of the Commune. It is curious to notice that, almost at the end, the
Girondists, frightened by the success of the Jacobins, tried to avert a
rising and to compromise with the King. Had Louis been willing to
replace them in office and to accept their terms unreservedly, it
seems almost certain that they would at the last moment have
broken with the Jacobins, and, in spite of their vaunted
republicanism, have ranged themselves on the King's side. But Louis
would listen to no overtures, and so they left him to his fate. It is also
curious to notice how small—even at this time of keen excitement
and revolutionary triumph—the insurrectionary battalions were, how
few voters supported them in the Sections, how the heads of the
party trembled for the result, and how difficult they found it to raise
an adequate and imposing force. Had the defenders of the Tuileries
had a leader to inspire them, had Louis for once laid sentiment aside
and displayed a flash of determined courage, the rising might have
been defeated and the mob dispersed. Napoleon, who watched the
whole scene from a window, and who afterwards declared in St.
Helena that the spectacle in the Tuileries gardens at nightfall was
more horrible than any of his battle-fields, believed that, had Louis
used his opportunities, he might easily have won the day. But the
irresolution, which had been his ruin, dogged the King's footsteps to
the last, and the 10th August ended in the capture of the Tuileries
and the destruction of the throne.
The six weeks which followed the victorious insurrection were weeks
of intense excitement in Paris. The two sections of Republicans
divided the spoil. Roland and his colleagues returned to office, and
Danton was appointed Minister of Justice. For the moment Danton
became the most conspicuous man in France. The young lawyer
had thrown himself into the revolutionary movement with
characteristic intrepidity and ardour. Reckless, cynical, unscrupulous
as he was, Danton bore the stamp of greatness. He was a king fit for
the turbulent, ambitious spirits, whom Robespierre was too timid a
theorist, and Marat too gloomy a fanatic to inspire. His physical
vigour, his stentorian voice, his eloquent fancy, his fierce contempt
for little men and little measures, the rough but genuine kindliness of
his nature, and his real enthusiasm for his country and for freedom,
appealed irresistibly to the imagination of his followers. Wherever he
had appeared, whether at the Cordeliers Club, in the early days of
the Revolution, or, later, among the officials of the Department of the
Seine, or, later still, in the Commune of Paris, he had made a
profound impression, and after the 10th August he naturally took the
lead. Unquestionably Danton had grave faults. He had too few
principles or scruples, little elevation of character, no refinement of
mind. But yet there is a certain air of grandeur round him. His
patriotism and courage cannot be doubted. His insight and capacity
for statesmanship stand in marked contrast to the incapacity of his
associates. No man learned more or more quickly from experience.
And of his surpassing force and influence there are a hundred
proofs.
In preparing the insurrection of the 10th August Danton had taken a
prominent part. He was rewarded by a high post in the Government,
and his action during the crisis which ensued is characteristic of the
man. He threw himself, heart and soul, into the national defence. He
felt that the one pressing necessity was to hold Paris against the
invaders. He knew that his party were a small minority, and that
Paris and France alike were full of men who would be ready enough
to turn against them and to compromise with the Allies. If France
were to be saved and the Revolution vindicated, he believed that
violent measures might be needed, and those violent measures he
was prepared to face. 'We must make,' he cried, 'the Royalists
afraid.' All the Jacobin leaders agreed in this. They saw that their
only chance of safety lay in paralysing their enemies by terror. But
some of them naturally hesitated as to the means. As the days went
on, the danger increased. On the 20th August, Lafayette, after a vain
attempt to induce his army to march on Paris, fled across the frontier.
The Allies rapidly advanced. On the 23rd August, Longwy, one of the
great frontier fortresses, surrendered with ignominious haste. On the
2nd September, Verdun surrendered too, and the road to Paris lay
open to the invaders.
In that time of terrible excitement, Danton and his colleagues carried
all before them. A special tribunal to try traitors was established on
Robespierre's demand. The property of the Emigrants was
confiscated. The refractory priests were condemned to
transportation. Urgent measures were taken to raise troops, and all
citizens, whether active or passive, were admitted to the National
Guard. Under the auspices of the Insurrectionary Commune, houses
were searched, arms seized, and suspected persons thrown into
prison. While the tocsin in the city sounded, Danton roused the spirit
of the Assembly. 'The alarm bell you hear rings no signal of danger.
It sounds the charge against the enemies of your country. To
conquer them, you must dare and dare and dare again, till France be
saved.' The Insurrectionary Commune, which still wielded the
powers which it had usurped on the 10th August, assumed dictatorial
authority, overawed the Ministers and the Assembly, and translated
Danton's warning into action. A frenzy of panic swept over Paris, and
the answer to the shameful surrender of Verdun was the famous
massacre in the prisons in September.
The exact responsibility for the massacre it is not easy to fix. Its
immediate cause was unquestionably the panic into which Paris was
thrown by the advance of the Allies. The terror of the moment
produced a civil war between those who felt that they were fighting
for their lives, and those who were supposed to be the friends of the
invaders. In that war the more desperate conquered, and the weaker
party fell. No doubt, the conquest was achieved by the aid of
ruffians. A crisis of extreme peril, when rumour is busy and suspicion
rife, brings to the surface many elements of disorder. The Ancien
Régime had left in Paris ample material for crime. It had taught the
poor to be ignorant and brutal. It had created a class of men to
whom pity and prosperity were equally unknown. Historians have
long disputed whether the Insurrectionary Commune was
responsible, or whether the movement was spontaneous. No doubt,
the leaders of the Commune, Panis, Sergent, Hébert, Billaud-
Varennes, and their guide and coadjutor Marat, gave it at least
encouragement and approval. It was Marat's policy triumphing at
last. But the truth is, Paris must share the responsibility, for during
those days of bloodshed, although the number of murderers was
very small, no one interfered. The Legislative Assembly looked on,
no doubt with grave compunction, but for all effective purposes with
indifference. The volunteers, the National Guards, the great body of
Parisian citizens stood by, apathetic or cowed. Robespierre may
have found it difficult to reconcile the massacre with his sentimental
love of virtue; but, though he was then a member of the Commune,
he did nothing to arrest its course. He could always persuade himself
in the last resort that murder was the justice of the people. Danton,
with sterner logic and audacity, believed it to be necessary to
paralyse the Royalists, and deliberately declined to interfere. Some
of the Girondists were horrified, and made honest attempts to check
the bloodshed, but attempts of an ineffectual kind. Others of the
Girondists seem to have regarded it as inevitable, if the Republic
were to be secured. Roland afterwards weakly tried to palliate it.
Pétion offered the murderers refreshment. The fact is that the
Girondist leaders were themselves exposed to the suspicious
hostility of the Commune, and were powerless or disinclined to act[9].
The outbreak in Paris was followed by similar outbreaks elsewhere.
But the sequel of the massacre showed the Jacobin leaders in a
better light. Under the guidance of Danton, a new spirit was infused
into the Government. Thousands of recruits poured into the French
camp, and the Assembly appointed to the chief command a soldier
of genius in the person of Dumouriez. The rapid success of the
invaders had been largely delusive. The fall of the French fortresses
was due less to the prowess of the enemy than to treachery among
the defenders. The Allied Powers were by no means at one upon all
points. Their generals and counsellors had neither brilliancy nor
dash. The Duke of Brunswick was already weary of the campaign.
Distinguished as he was, he belonged to an old school of soldiers,
and he had never wished to march direct on Paris, leaving his
communications unprotected. He had little sympathy with the
Emigrants. He was in bad health and half-hearted in the war. The
Austrians had not supported him as strongly as they had promised to
do. His troops had suffered severely from illness and bad weather.
Even after the fall of Verdun, he was disinclined to persevere.
Accordingly, when Dumouriez gathered his forces at Valmy, and
risking an engagement on the 20th September, succeeding in
checking the enemy's advance, the Duke took the opportunity to
retire, the campaign was abandoned, and Paris was saved.
From Valmy the tide of victory rolled on. Dumouriez followed up his
success by the battle of Jemappes and the conquest of Belgium. But
in the spring of 1793, fortune turned against him, and Dumouriez,
like Lafayette, threw up his post and fled. Once again, in the summer
of 1793, the Allies threatened to advance on Paris. Once again the
fear of invasion strengthened the hands of the Terrorist party. Once
again a new general was discovered, and the vigour of the Jacobins
carried the day. Once again the young Republic triumphed, and after
the winter of 1793 all danger from the frontiers disappeared. The war
assumed a new character. It became a war of propaganda, and
swept over Europe. It found upon its borders an ancient society,
already in a state of dissolution, which, devoid of patriotism,
enthusiasm or popularity, fell to pieces before its attack. Then,
forgetting its philanthropic principles, it returned to the practices of
the past, and in its dealings with Europe, adopted the spirit of the
system which it assailed. In the end, the French Revolutionary war
did not abolish tyrants or unite the human race. It obliterated many
old landmarks. It broke down many feudal barriers. It swept many
little despotisms away. It drew more sharply the divisions between
the different nations. It rallied them more closely round their
sovereigns. It cleared the ground for modern Europe to grow up.
Unconsciously, and cruelly, it laid the foundations of that united
German Empire, which was one day to take a terrible revenge on
France. But the old order, thus improved and altered, was too strong
to die. Ultimately it subdued its conquerors, and the hero of the
victorious Revolution married the niece of Marie Antoinette.

FOOTNOTES:
[8] Some contemporary writers estimate the numbers of Condé's
army very low, at not more than 3000 or 4000. M. Sorel, who is
always a valuable authority, reckons them at 10,000 in October,
1791. Mr. Morse Stephens gives a much higher figure.
[9] It cannot be proved decisively that the Insurrectionary
Commune organised or paid for the massacre, but several facts
seem to point in that direction. The subject is fully discussed in M.
Mortimer-Ternaux's Histoire de la Terreur, vol. iii. Appendix XVIII,
and by M. Louis Blanc, in the note to the second chapter of the
eighth book of his history.

CHAPTER VIII.
The Fall of the Gironde.
The first sitting of the National Convention was held on the 21st
September, 1792. The Parisian deputies, elected two or three weeks
before, under the eyes of the Insurrectionary Commune, belonged
almost entirely to the Jacobin party. Their election is a significant
example of the methods of Jacobin organisation, and in that election
Robespierre, always a vigilant wire-puller, had shown the adroitness
of his tactics and had taken the most prominent part. The meeting of
the electors took place on the 2nd September, the day when the
prison massacres began. No sooner had the electors assembled,
than they were transferred, by the directions of the Commune, from
their ordinary meeting-place to the Jacobin Club. Some of the
electors, who held moderate opinions, were then excluded by a
preliminary vote. The system of secret voting was suspended, and
all were obliged to vote openly before an audience loudly and
watchfully alert. As one Jacobin speaker admitted, his party would
have been 'beaten, even in Paris, in any election in which the voting
had been secret.' The result was the victory of the extreme
politicians. Robespierre, Danton, Marat and Collot d'Herbois were
among the best known deputies elected, and all the heroes of the
Insurrectionary Commune were brought triumphantly into the
Convention.
In the provinces, however, the Jacobins were less successful. There
too the minority carried the day, and most of the electors stayed
away from the polls. But still the minority in the provinces
represented a considerable bulk of opinion. Everywhere the result of
the elections was to confirm the Coup d'etat of the 10th August. All
who still took part in politics seemed to realise that the cause of the
King was incompatible with the defence of the country, and preferred
to put the interests of the country first. The consequence was that
the Girondist party had a large following in the new Assembly. The
Jacobins, it is true, were more compact and vigorous, and the great
majority of members in the centre had no very definite views, and
could not be depended on to vote consistently. But still the Girondist
position was strong. They had the command of the Government.
They had eloquent and effective speakers. They had several men of
character and ability. They had behind them the weight of moderate
opinion, which was shocked by the fearful disorders of September.
The massacres had produced a reaction which tended, now that the
danger of invasion was over, to strengthen the Girondist ranks. Had
the Girondists possessed any organisation, any instincts of party
management, or any leaders of authority and insight, they might
have formed a powerful party, and have guided the Revolution yet.
But unfortunately the Girondists had none of these things.
Vergniaud, their most splendid orator, had none of the qualities
needed for a leader. Guadet and Gensonné, the two brilliant
advocates who accompanied him from Bordeaux, were no better
able to guide a party. Pétion, the ex-mayor, had proved his
incompetence already. Roland, with all his honesty and aspirations
after order, had little real capacity or strength. Condorcet, the
philosopher of the Gironde, brought to the pursuit of politics all the
characteristic vices of the academic mind. Barbaroux, the hero of the
Marseillais, was only distinguished by his beauty from the rest.
Isnard and Louvet, Lanjuinais and Gorsas could not supply what was
wanting in their colleagues. The party itself had no cohesion. Brissot,
who had for some time been its leader, could not impose his
ascendency for long, and found his authority challenged by the rising
reputation of Buzot. Gradually two sections of Brissotins and
Buzotins grew up within the ranks of the Gironde, and rendered still
remoter than before the prospects of united and decisive action on
the part of the majority of the House.
Buzot owed to Madame Roland much of the influence which he
enjoyed with his party. Her house was the chosen resort of the
Gironde. Their policy was largely arranged in her drawing-room. Her
husband was their chief representative in the Government, and her
interest in her husband's policy was as well known as her
attachment to Buzot. Madame Roland is known to us by the portrait
which she herself has drawn, and that portrait shows us clearly her
undoubted courage and ability, her enthusiasm for the philosophy
which she had studied and for the ideals which her bright
imagination loved. But the memoirs show us also the self-conscious
genius of the writer, her swift but rather shallow judgments, the
strong personal element in her opinions, the ill-controlled, ambitious
restlessness of her generous and ardent mind, and her incapacity for
moderation, for being just towards opponents or tender towards
fallen foes. What part Madame Roland played in politics it is
impossible to say exactly. We know something of her words and
actions upon a few occasions, and those, in spite of the charm and
romance which surround her, are not always creditable to her head
or heart. She died bravely, and posterity, recognising that, has
perhaps been bountiful towards her virtues. But in so far as she
inspired the Girondists, her political influence can only be regarded
as disastrous, for there never was a party worse advised.
Other women, fair and unfair, lovely and unloveable, appear in the
story of the French Revolution. There were the salons of the early
days, where the Royalists gathered at the houses of Madame de
Chambonas and Madame de Sabran, while the other side found
more congenial company in the rooms of Mesdames de Beauharnais
and Talma. There was the salon of Madame de Genlis with its
traditions of Orleanist intrigue, the official society of Madame
Necker's circle, and Madame de Broglie's coterie of young, well-bred
reformers. Later on, there was the salon of Madame de Staël, where
the accomplished hostess pushed the interests of Narbonne as
devotedly as Madame Roland pushed the interests of Buzot, and the
well-lit tables of Madame de Sainte-Amaranthe, who expected to
profit by the guests whom she received. There were the quieter but
happier homes of Madame de Condorcet and Lucile Desmoulins.
There were the not less happy women whom Danton, Robespierre
and St. Just loved. There was the Queen, always gallant and
unfortunate, but in her political influence most unfortunate of all;
Charlotte Corday, the Girondist avenger, whose enthusiasm veiled
from her the ugliness of murder; Theresa Tallien, who gave up her
life and honour to a worthless man and used her power afterwards
for purposes of mercy; Olympe de Gouges, the hapless dramatist
and pamphleteer, who played her triple part as hostess, celebrity and
victim in the Terror; and many another notable woman, of reputation
high or low, of influence worthy or unworthy, such as the Demoiselles
de Fernig, who served as aides-de-camp in Dumouriez' camp,
Théroigne de Méricourt, known by all and loved by many, who,
armed with a naked sabre, led the wild women in October to
Versailles, Sophie Momoro, who headed the orgies on the Feast of
Reason, and Rose Lacombe, the queen of the vile women who
haunted the clubs and streets and galleries, disgraced the National
Assembly, and knitted round the guillotine. It is strange, but it would
seem to be indisputable, that, in many of the worst excesses of the
Revolution, women of the worst character were specially
conspicuous, and that when politics degenerated into savagery
women without womanliness excelled the most.
The nine months which followed the meeting of the Convention were
occupied by the struggle between the Jacobins and the Gironde.
Even before the 10th August that struggle had begun. After the fall of
Louis, it had become acute; and as soon as the Convention met, the
animosities of the rival parties blazed out afresh. The difference in
principle between Girondists and Jacobins is not altogether easy to
define. Many tenets they professed to share in common, and until
the 10th August their aims had been in some respects the same.
One noticeable difference, however, between them, lies in the
character of the men. The Jacobins, as a whole, though the rule is
subject to exceptions, were men of a rougher stamp, occasionally ill-
educated, coarse and unscrupulous, in some cases cruel, in many
cases corrupt, but practical, alert and intrepid politicians, prepared to
run great risks, merciless to their enemies if they conquered,
sometimes ready with gay desperation to suffer if they lost the game.
The Girondists, as a whole, were men of higher intellectual calibre,
of more education and refinement, of a better cast. They were
honest and decent. Their intentions were pure. They were
embarrassed by scruples in a struggle with opponents who had
none. They were distinguished by a fine enthusiasm for their vague
and delusive ideals, which, if it seems theatrical now, was genuine
then, and to which, while lamenting its waste and its errors, one may
pay the tribute of respect.
But when one comes to principles, the difference between the two
parties is less clear. The truth is that in the French Revolution there
was no place for the Gironde. Two parties, and only two, were
possible. One was the party which, representing the great majority of
French feeling, had made the Revolution, had swept away the
Ancien Régime, had founded the first constitution and had taken its
name from its creation, the party which had aimed at establishing
political freedom and a new system based on that in France. This
party cared for liberty and order but wished for little more, and its
mistakes and the fortunes of the time gradually lost it the control of
events. The other was the new party, which rose to power on its
rival's faults, which cared little for liberty and less for order, but which
hoped to use the forces of distress and discontent to grasp the
power which its rival had monopolised, to found a new social and
industrial system in which it would secure a fairer profit for itself, and
to destroy without scruple or compassion all who impeded the
realisation of its aims.
Between these two parties the Girondists stood. They belonged to
neither, but they shared to some extent the views of both. Like the
first, they had moderation, a sense of restraint and a love of order.
Like the second, they repudiated any compromise with the past and
hoped to establish their Utopia. But though they were strong enough
to defeat the first, they were not strong enough to resist the second,
and they could not fuse with either. They could not join with the
Constitutional party, or rally the moderate majority round them,
because, though they shared its feeling of propriety, they scorned its
tenets and prejudices, its king and its religion. They could not throw
themselves into the arms of the Jacobins, because, though they
liked some of their democratic schemes, they could not countenance
the Jacobin excesses or the Jacobin intrigues, the ruthless levelling
of the Jacobin maxims, the Jacobin contempt for property and life.
Nor could the Girondists found a party of their own. Their ideas were
not sufficiently definite for that, nor sufficiently different from those of
others. The only principles which belonged distinctively to them were
an enthusiasm for the forms of a republic, and an enthusiasm for the
policy of war. The latter was partly an ill-considered emotion and
partly a tactical device. The former was not a principle at all. The
essence of republicanism, which is government by the people, had
been accepted by all parties except the most reactionary, and was
not peculiar to the Girondist belief. The forms of it, which, possessed
by an extraordinary desire to emulate the Greeks and Romans, the
Girondists esteemed so highly, were hardly worth a struggle to
obtain. The fact that the Girondists should have cared for republican
forms so much more than for anything else, is perhaps the clearest
proof of their incompetence as practical politicians. For, as
Robespierre had the sense to see, the term republic is an empty
name, which the faith and heroism of men have sometimes
associated with ideals of purity and freedom, but which has often
been only a disguise for governments that were neither free nor
pure.
As soon as the Convention opened, the Republic was proclaimed,
and the struggle between the rival parties began. For the moment
the Girondists were the stronger, and they were determined to use
their power to suppress and, if they could, to punish the leaders of
the Insurrectionary Commune. In that attempt they were partially
successful. The steady persistence of Roland, supported by the
majority of the Convention, succeeded at last in dissolving the
Insurrectionary Commune. On the 30th November, a moderate
politician, Chambon, was elected Mayor by a large majority, after two
other moderates, Pétion and d'Ormesson, had previously been
elected and had refused to serve. The council and the other officials
of the Commune were also renewed. On this occasion the voting
was secret; and although the Jacobins spared no efforts, and
succeeded in carrying Hébert and Chaumette, the smallness of the
Jacobin vote and the abstention of the vast majority of voters
showed how weak numerically were the forces which the minority
could command.
But in other respects the Girondists were less successful. Their
proposals for the formation of a guard for the Convention resulted
only in bringing to Paris a small force of Fédérés from the
departments. Their demand for the punishment of those concerned
in the September massacres fell to the ground. Their attacks upon
Robespierre, Marat and others produced only bitter personalities,
which tended to weary the Assembly, and by giving Robespierre
opportunities of dilating on his services to the Revolution, to increase
his popularity in Paris. The Jacobins began to threaten their
monopoly of office. In the middle of October, Pache, the newly
appointed Minister of War, and till then an intimate friend of Roland,
cast himself into the arms of the Jacobin party. Suddenly turning on
his Girondist colleagues, Pache made the War Office the meeting-
place of the politicians of the Commune, placed his influence and his
funds at their disposal, and to the disgust of Dumouriez and Danton,
threw the military arrangements into confusion.
The Girondists were further weakened by the trial of the King. The
long debates upon that question, which began early in November,
1792, and which ended in Louis' execution on the 21st January,
1793, certainly damaged the reputation of the party. They showed in
a clear light the stern logic of the Jacobin leaders and the weakness
and disunion of their opponents. They gave fresh opportunities for
excitement and disorder, which the Jacobins knew how to use. The
vote which condemned the King to death was carried finally by a
narrow majority, but it could not have been carried without wholesale
intimidation. The violence of the agitation in the galleries, in the
streets, in the Sections, which steadily rose as the debates went
forward and as a feeling of sympathy for Louis appeared, produced
so general a panic, that it is recorded that fourteen thousand people
fled from Paris in the last week of the year, under the impression that
the massacres of September were about to be repeated[10]. It is true
that in the end the leading Girondists voted for death; but they voted
openly in the presence of an armed and vociferating crowd, amply
sufficient to decide the wavering and almost sufficient to terrorise the
brave. Vergniaud, who voted with the majority when the critical
moment came, had already pleaded for mercy in the finest speech
which he ever delivered, and had declared the night before the
verdict that it was an insult to suppose him capable of voting for
Louis' death.
With the trial of the King the demoralisation of politics increased. The
Convention lost all dignity and decorum. The violence of the rival
parties deepened. Deputies came down armed to the meetings of
the House. The president, powerless to keep order, was frequently
insulted in the chair. Abusive terms were shouted across the floor.
The voices of the speakers were constantly drowned in the din from
the galleries, where, according to Brissot, 'the brigands and
bacchantes' ruled. The same demoralisation appeared in the public
service. On the proclamation of the Republic, in September, 1792, all
administrative and judicial officers were renewed. The Convention
declared that a knowledge of the law should no longer be a
necessary qualification for judicial appointments. Education was
regarded as equally unnecessary, and a number of ignorant and
incompetent officials were thus imported into the administration. The
multiplication of offices and places, so profitable to those in power,
rapidly increased as time went on, and with the spirit of plunder the
spirit of corruption spread. Under Pache the War-Office became a
centre of Jacobin intrigue, where the Minister and his associates
could display with effect their bitter distrust of the Commander-in-
Chief, undermine Dumouriez' authority in the army, and, regardless
of his wishes and designs, promote their own theories and provide
for their friends.
From the beginning of the new year the Girondists steadily lost
ground. In January, Roland, their most active supporter in the
Government, resigned his office. The control of the Ministry of the
Interior, with all its authority and resources, thereupon passed into
the hands of Garat, a man of amiable intentions and moderate
views, but entirely lacking in force or decision, and with none of
Roland's devotion to the Gironde. Early in February, Pache, who had
been compelled to retire from the War Office, to the delight of
Dumouriez and Danton, was elected Mayor in Chambon's place, and
in his person the Jacobins finally regained control of the Commune
of Paris. About the same time Condorcet brought forward the
Girondist proposals for a new constitution, proposals wildly
unpractical in their nature, which gave satisfaction to no one at all,
and which lent some colour to the charge, which the Jacobins
pressed against the Gironde, that they wished to confer powers upon
the departments which would make them almost independent
States, to destroy the influence of the Government in Paris, and to
break up the unity of the Republic. The Girondists, who had no large
following in the capital, proceeded to alienate what following they
had. They declared irreconcilable war upon the Commune. They
denounced the disorder of the Parisian mobs, and their demands for
exceptional legislation in their favour. They boasted unwisely of the
devotion of the provinces to themselves. They threatened to punish
heavily any attempt at intimidation by the Sections, but they took no
steps to guard effectually against it. Finally, they made an attack
upon Danton as ill-judged as it was unprovoked, and thus alienated
the only man who had influence and ability enough to save them,
and who, weary of factious animosities and earnestly desiring to
found a Government strong enough to make itself respected, might
with a little tact have been induced to offer them his powerful
support.
Moreover, the course of external politics once again assisted the
Jacobin designs. The victory of Valmy had been followed by a series
of successes on the Rhine, in Savoy and in Nice, by the defeat of the
Austrians and the conquest of Belgium. But the reckless policy of the
Convention, its disregard of treaties, and its determination to spread
revolutionary principles at any cost, multiplied the enemies of
France. The French Government's resolution to attack Holland
offended and alarmed the English. The execution of Louis created
deep and general indignation in Europe. Early in 1793, England,
unheeding Pitt's pacific dreams, and roused by the warning tones of
Burke's hot anger and imagination, plunged into the war. Spain,
under its Bourbon princes, followed suit. The difficulties of the
French troops increased as their spirit and discipline diminished. The
allied armies resumed the offensive. At the beginning of March a
succession of reverses overtook the French arms, and the invasion
of Holland was abandoned. On the 18th, Dumouriez with the main
army was defeated by the Austrians in the battle of Neerwinden, and
Belgium was lost. Dumouriez, disappointed by the turn of events,
long weary of the Jacobin ascendency and meditating means to
overthrow it, rejected Danton's friendly encouragement, talked
openly of restoring the Constitutional throne, and determined to
declare against the Convention. The Convention, aware of his
designs, sent off commissioners to arrest him in his camp. On the
3rd April, foiled at the last by his own irresolution and by the apathy
of his troops, Dumouriez left his army and took refuge in the Austrian
ranks. Once again the French commander had deserted in the face
of disaster, and the danger of invasion reappeared.
The events on the frontier reacted immediately upon politics in Paris.
Danton at last succeeded in convincing the Assembly of the absolute
necessity of a strong Executive. In the end of March and the
beginning of April, a series of decrees passed the Convention,
establishing, for the first time since the outbreak of the Revolution, a
powerful Government in France, and founding or re-organising at the
same time the three chief instruments of the Terror. One decree
created the Revolutionary Tribunal, a court with summary process
and extraordinary powers, to try conspirators against the State;
another, the famous Committee of General Security, to hunt down
and punish political crime; and a third, the still more famous
Committee of Public Safety, soon to become the most redoubtable
despotism in the history of the world. With these decrees went other
energetic measures—fresh powers for the commissioners, the
'Representatives on Missions,' sent into the provinces to execute the
orders of the Convention; a fresh levy of 300,000 men for the